MASTER THESIS - UiS Brage

Faculty of Science and Technology

MASTER’S THESIS

Study program/ Specialization:

Offshore Technology/ Industrial Asset

Management

Spring semester, 2015

Open / Restricted access

Writer:

Ilya Murashov

………………………………………… (Writer’s signature)

Faculty supervisor: Prof. Ove T. Gudmestad

External supervisor(s): Dr. Najam Beg (Caltec Ltd.)

Thesis title:

Severe slugging phenomenon and a novel method for its mitigation based on the Surface Jet

Pump technology.

Credits (ECTS): 30

Key words:

Severe slugging, OLGA, mitigation

techniques, surface jet pump.

Pages: 93

+ enclosure: 9 pages

Stavanger, 29/06/2015

Date/year

i

Abstract The present thesis is focused on the problem of severe slugging and ways to mitigate it. Severe

slugging is an oscillatory multiphase flow regime characterized by high variations in

production rates occurring in offshore pipeline-riser systems.

Chapter 1 provides basic notions related to multiphase flow, which are essential for

understanding of the rest of the thesis.

Chapter 2 gives a thorough description of the severe slugging occurrence mechanism and

preconditions as well as introduces different types of the phenomenon. Special attention is

given to the effect of mass transfer and how it alters the flow regime’s behavior. Detrimental

effects of severe slugging are discussed and some examples are provided.

Chapter 3, making a significant part of the thesis, provides its reader with carefully gathered

data concerning severe slugging alleviation and mitigation methods published from 1973 to

2015, both conventional and purely speculative methods are discussed. Examples, where

possible, are given.

Chapter 4 considers modeling of severe slugging in a vertical riser with aids of the multiphase

simulation program OLGA. A constructed study case is considered and described with some

of the mitigation techniques implemented and tested.

Chapter 5 evaluates a novel severe slugging mitigation method proposed by Caltec Ltd. UK.

The method assumes pipeline system depressurization by installation of a Surface Jet Pump on

the production platform. The chapter gives the method description and verifies its feasibility

using a simulation model within OLGA.

The thesis ends with Conclusions and Recommendations for further work and self-evaluation.

ii

Acknowledgments

I would like to express my appreciation to my supervisor at the University Prof. Ove. T.

Gudmestad for his thorough support. This man is one of the most dedicated educators I’ve ever

met who spends an immense amount of time to this cause. I am also thankful to Prof. Jayantha

P. Liyanage, the head of Industrial Asset Management program, which provided me with this

wonderful chance to study at the University of Stavanger.

Thanks are due to Dr. Najam Beg, Technology Director and co-founder of Caltec Ltd., for him

to introduce me to the jet pump technology through his passionate presentations and personal

conversations, and his colleague at Caltec Dr. Rhodri Jones for guidance throughout the thesis.

My sincere gratitude goes to Prof. Conrad Carstensen for dedicating some of his leisure time

to lengthy conversations over the phone explaining process nuances I was puzzled with.

I am thankful to Prof. Aly A. Hamouda for providing me with an academic version of OLGA

and PhD candidate Milad Khatibi for helping me with initial steps in this software. My

gratitude also addressed to Egor Shevchenko (Schlumberger) for spending his free time to

verify my OLGA models.

Finally, I’d like to thank my friends in Stavanger, especially Arkhat Sultabayev, who helped

resolving my questions in production and reservoir aspects.

iii

Nomenclature 𝐴 – cross sectional area, [m2];

𝐶 – choke coefficient [Pa∙s2∙m-2];

𝐷 – diameter, [m];

𝐹 – force per area, [N∙m-2, Pa];

𝑀 – molar mass [g∙mol-1];

𝑈 – phase velocity, [m∙s-1];

𝑃 – pressure, [Pa];

𝑄 – volumetric flowrate, [m3∙s-1];

𝑅 – gas constant [=8314 J∙K-1∙kmol];

𝑇 – temperature [K];

𝑓 – friction coefficient [-];

ℎ – height, [m];

𝑙 – length, [m];

�̇� – mass flow rate [kg∙s-1];

𝑧 – compressibility factor, [-];

𝛼 – phase area fraction, [-];

𝛽 – pipeline inclination angle to horizontal line, [rad];

𝛾 – specific gravity, [-];

𝜌 – density, [kg∙m-3];

𝜙 – average holdup, [-];

iv

Subscripts

𝐵 – backpressure;

𝐺 – relating to gas;

𝐻𝑃 – high pressure;

𝐿 – relating to liquid;

𝐿𝑃 – low pressure;

𝑀 – relating to mixture;

𝑀𝑃 – medium pressure;

𝑂 – relating to oil;

𝑅 – relating to riser;

𝑃 – relating to pipeline;

𝑏 – relating to bubble-point;

𝑏𝑢𝑏 – relating to bubble;

𝑟𝑒𝑠 – relating to reservoir conditions;

𝑠𝑒𝑝 – relating to separator;

𝑤𝑓 – relating to flowing bottom hole pressure;

0 – at standard conditions;

Abbreviations

𝐴𝑃𝐼 – American Petroleum Institute;

𝐶 − 𝐸𝑆𝑃 – Caisson – Electrical Submersible Pump;

𝐸𝑃𝐼 – elimination performance index;

𝐸𝑆𝑃 – electrical submersible pump;

𝐹𝑉𝐹 – formation volume factor;

v

𝐺𝑂𝑅 – gas-oil ratio;

𝐻𝑃 – high pressure;

𝐼𝐷 – internal diameter;

𝐼𝑃𝑅 – inflow performance relationship;

𝐼 − 𝑆𝐸𝑃 – compact cyclonic gas/liquid separator by Caltec Ltd.;

𝐿𝑃 – low pressure;

𝑀𝑃 – medium pressure;

𝑂𝐿𝐺𝐴 – (stands for OiL and GAs) multiphase flow simulator;

𝑃𝐼 – productivity index;

𝑃𝑉𝑇 – pressure-volume-temperature;

𝑅𝑆𝐺𝑂 – solution gas-oil ratio (the same as 𝑅𝑆);

𝑆𝐽𝑃 – surface jet pump;

𝑆𝑆 – severe slugging;

𝑆𝑇𝐵 – barrel at standard conditions;

𝑆3 – Slug Suppression System;

𝑉𝐴𝑆𝑃𝑆 – Vertical Annular Separation and Pumping System;

vi

Contents

Abstract ....................................................................................................................................... i

Nomenclature ........................................................................................................................... iii

Contents .................................................................................................................................... vi

Introduction ................................................................................................................................ 1

1. Background ............................................................................................................................ 2

1.1. Multiphase flow.............................................................................................................. 2

1.2. Slug flow ........................................................................................................................ 4

2. Severe slugging phenomenon ................................................................................................ 6

2.1. Mechanism ..................................................................................................................... 6

2.2. Severe slugging classification ...................................................................................... 10

2.3. Mass transfer effect on severe slugging ....................................................................... 11

2.4. Severe slugging occurrence criteria ............................................................................. 15

2.4.1. Stratified flow criterion ........................................................................................ 15

2.4.2. Bøe criterion ......................................................................................................... 16

2.4.3. Pots criterion ......................................................................................................... 16

2.4.4. Taitel criterion ...................................................................................................... 17

2.5. Effect on production system ......................................................................................... 18

3. Severe slugging elimination ................................................................................................. 20

3.1. Conventional methods .................................................................................................. 20

3.1.1. Backpressure increase........................................................................................... 20

3.1.2. Choking ................................................................................................................ 20

3.1.3. Riser-base gas injection ........................................................................................ 24

3.1.4. Active control methods ......................................................................................... 25

3.2. Unconventional methods .............................................................................................. 26

3.2.1. Self-lift .................................................................................................................. 26

3.2.2. A conduit of reduced diameter ............................................................................. 31

3.2.3. Flow conditioning, phase agitation, mixing ......................................................... 32

3.2.4. Separation ............................................................................................................. 37

3.2.5. Foaming ................................................................................................................ 42

3.2.6. Multiphase riser base lift ...................................................................................... 48

4. Mitigation methods modelling ............................................................................................. 50

4.1. Case study construction ................................................................................................ 50

vii

4.1.1. The pipeline-riser system...................................................................................... 50

4.1.2. The fluid model .................................................................................................... 51

4.1.3. The well and wellbore .......................................................................................... 54

4.1.4. The topside facilities ............................................................................................. 54

4.1.5. Solution convergence analysis .............................................................................. 55

4.1.6. Flow regime map identification ............................................................................ 58

4.2. Topside choking ........................................................................................................... 60

4.3. Riser-base gas injection ................................................................................................ 66

4.4. Diameter reduction ....................................................................................................... 68

5. Severe slugging mitigation by backpressure reduction with a jet pump .............................. 69

5.1. Proposal description ..................................................................................................... 69

5.2. Hypothesis .................................................................................................................... 70

5.3. Jet pump technology..................................................................................................... 71

5.3.1. History and applications ....................................................................................... 71

5.3.2. Principle of operation ........................................................................................... 72

5.3.3. Typical construction ............................................................................................. 73

5.3.4. Applications in oil and gas industry ..................................................................... 75

5.4. Effect on a constant mass flowrate system................................................................... 77

5.5. Effect on a system with producing well ....................................................................... 79

5.6. Jet pump calculation example ...................................................................................... 82

Conclusions .............................................................................................................................. 85

Recommendations for further work, self-evaluation ............................................................... 87

Reference list ........................................................................................................................... 88

Appendix A. ............................................................................................................................. 94

Appendix B. ............................................................................................................................. 96

Appendix C. ............................................................................................................................. 98

Appendix D. ........................................................................................................................... 100

Appendix E. ........................................................................................................................... 102

1

Introduction Nowadays, a significant amount of hydrocarbons is being produced from marginal fields, which

development assumes subsea production followed by a flowline terminated into a platform or

onshore production facilities via a riser. Because of the challenges related to subsea processing,

production fluids are often transported in a multiphase flow, which imposes additional

requirements on flow assurance, such as oscillatory flow regimes.

The problem of severe slugging became evident in early 70s due to both an increased number

of marginal fields’ developments and increased water depths. Severe slugging may arise when

a number of preconditions are present, such as multiphase flow at relatively low production

rates, and the specifics of the pipeline-riser geometry.

Severe slugging can be characterized by low-frequency oscillatory flow with significant

pressure and flowrate fluctuations. It has proved to be hazardous for the production system in

general, potentially causing high stress loads on riser pipework and topside piping and affecting

efficiency of process equipment, such as separators, pumps and compressors. If untreated,

severe slugging may result in complete system shut-down. In addition to that, the phenomenon

increases backpressure imposed on the wells, decreasing production by a substantial amount or

killing the wells completely in some extreme examples.

A number of mitigation techniques have been proposed over the last 40 years. The thesis gives

a thorough description of all the methods found in literature ranging from 1973 to present and

evaluates on a novel mitigation technique proposed by Caltec Ltd., UK.

2

1. Background The purpose of this chapter is to introduce some of the basic concepts important for the thesis

topic. In case the reader is familiar with the definitions, the chapter may be skipped without loss

of understanding.

1.1. Multiphase flow

Multiphase flow in pipeline simply refers to flow regimes, consisting of more than one phase.

For petroleum multiphase pipeline transportation the most commonly encountered fluids are

different fractions of hydrocarbons (gas, oil, gas condensate), produced water, and solid

particles. Therefore, a typical case when gas, oil, and water are present in the conduit refers to

three-phase flow.

Unlike single phase flow, multiphase flow may appear in different forms depending on the

conduit geometry and the fluids properties and quantities. These forms are called flow regimes.

Figure 1.1 gives an example of a flow regimes map of two immiscible fluids plotted with respect

to gas and liquid superficial velocities. The superficial velocity is a multiphase flow specific

definition and for a given flow phase may be defined as a ratio between the phase volumetric

flowrate and the pipe cross section area:

𝑈𝑆𝑖 =𝑄𝑖𝐴. (1.1)

The superficial velocity definition above provides an intuitive understanding of the term:

superficial velocity of a phase in a multiphase flow is the average instantaneous velocity of the

given phase if it was a single phase flow with the given volumetric flowrate. On the other hand,

the volumetric flow rate of a given phase may be expressed in terms of the absolute phase

velocity (𝑈𝑖) and the area occupied by the phase (𝛼𝑖𝐴):

𝑄𝑖 = 𝛼𝑖𝐴𝑈𝑖 , (1.2)

where 𝛼𝑖 = 𝐴𝑖 𝐴⁄ is the area (volume) fraction of phase 𝑖.

When referred to liquids, area fraction is commonly called (liquid) holdup. For petroleum

related topics, water phase fraction is often referred to as water cut.

3

Thus, we can describe superficial velocity of phase 𝑖 in terms of its absolute velocity in the flow

and area fraction:

𝑈𝑆𝑖 = 𝛼𝑖𝑈𝑖. (1.3)

Some of the flow regimes observed in horizontal pipes essential for the thesis are to be explained

here. Starting with modest amount of liquid and gas in the pipe, the stratified flow may be

observed, refer to Figure 1.1. Substantial increase of gas quantities flowing through the pipe

induce waves on the surface of the liquid phase forming stratified wavy flow up to the point

where the wave crests reach the top of the pipe and block the whole cross section of the conduit

thus forming relatively short liquid slugs. Such phenomenon is called hydrodynamic slugging.

Further increase of gas velocities creates sufficient turbulence to tear the liquid droplets of the

liquid phase, forming a layer of liquid around the inner circumference of the pipe and dispersed

Figure 1.1. An example of flow regime map for two-phase gas-liquid flow in a horizontal

pipe [1].

The figure uses nomenclature different from the rest of the thesis.

4

liquid droplets traveling with the gas. This flow regime is called annular flow and frequently

encountered in gas with gas-condensate flows. On the other extreme, when liquid superficial

velocity is high enough, the gas phase is dispersed in liquid and dispersed-bubble flow occurs.

Between the phases mentioned above, a number of intermittent flow regimes lay, characterized

by less even distribution of liquid and gaseous phases. To those slug and churn flows may be

attributed. The overall flow regimes classification is not limited to those mentioned in this

paragraph and not universal across the field. It may be encountered that individual researchers

as well as publications of different years tend to specify flow regimes in a slightly different way.

However, flow maps are restricted in their usefulness due to predefined fluids properties and

pipeline geometry. They are also of limited use for more-than-two-phase flows. The situation is

further worsened if the fluids are able to have mass transfer between them, which is often the

case for petroleum pipelines. Thus, depending on PVT conditions and oil properties, gas is

dissolvable in oil phase; water may be present in form of vapor in the gas phase, and so on.

1.2. Slug flow

Slug flow is characterized by varying liquid and gas flowrates and may impose significant threat

to production facilities, inducing oscillations, increasing wear and tear of equipment, hindering

separation efficiency. Severity of slug flow mainly depends on its origin and may be described

in terms of liquid slug length and pressure fluctuations.

Three main types of slugging by its origin related to petroleum multiphase production are:

1. Hydrodynamic slugging. As was described before, this type of slugging is formed due

to waves generated by gas blowing with sufficient speeds over the liquid phase layer.

When the wave crests reach the top of the pipe, they block the whole cross section of the

conduit and form slugs. As a rule of thumb, slugs generated by this mechanism are

relatively short.

2. Operationally induced slugging. Refers to transient flow regimes in multiphase

pipeline, such as production shut-down/ramp-up/restart and pigging operations.

3. Terrain induced slugging. Refers to slug flow originated at the dips of pipelines

following the profile of the terrain (therefore the name). Liquid accumulated at a dip

may block the pipeline cross section and form a slug. If a number of criteria met, the

5

slug may grow up to considerable lengths until the upstream gas pressure build-up

pushes it out of the dip.

An extreme case of terrain induced slugging often occurs in pipeline-riser systems and called

riser (induced) slugging, or severe slugging. Even though severe slugging is attributed to a

separate type of slugging by some authors, both have the same mechanism and only differ in

the location of the origin.

6

2. Severe slugging phenomenon

2.1. Mechanism

Severe slugging as a separate phenomenon received much attention in early 80s as part of the

Tulsa University Fluid Flow Projects run by Schmidt [2-4] and was probably linked to higher

number of subsea developments and increased depths of offshore developments, therefore taller

riser and severer riser induced slugging conditions. In a series of studies conducted for the

project, two main flow regimes associated with slug flow in risers were distinguished. First, the

hydrodynamic slugging, originated in upstream pipe and arriving to the riser base and moving

up the riser conduit. The second one, severe slugging, was classified as terrain induced slugging

due to its nature, was governed by different mechanism and had severer consequences.

Severe slugging was defined as such a terrain induced slug flow regime resulting in slugs of

length equal or greater than that of the riser. It was also found to be a cause of high magnitude

pressure fluctuations and varying production characteristics, both in terms of production rates

and phases delivered to the platform (see Figure 2.1). First precondition for severe slugging

occurrence was found to be low gas and liquid flowrates [3].

7

Severe slugging mechanism is initiated by formation of liquid blockage at the riser base due to

a low point in the pipeline, which prevents the gas phase from penetrating inside of the riser,

and followed by force balance between the increasing hydrostatic pressure of the liquid column

PR

QL

QG

I II III+IV

t

t

t

(I) Liquid buildup

(II) Slug production

(III) Bubble penetration

(IV) Gas blowdown followed by

liquid fallback

Figure 2.1. Qualitative depiction of one severe slugging cycle.

Figure a) represents variations of riser base pressure, liquid and gas flowrates with time.

Figure b) shows liquid holdup distribution along the system (adopted from [5]).

a) b)

8

in the riser and compressed gas pressure inside of the pipeline. The whole process may be broken

down on four major stages (see Figure 2.1):

I. Liquid buildup

Liquid accumulated at the riser base blocks the conduit cross section and creates pressure

difference between upstream and downstream of its origin. Often initiated by the

preceding stage of liquid fall back, when some part of the slug produced during the

blowout phase loses momentum and falls to the riser base. The liquid phase, continuing

flowing to the riser base in stratified flow, initially increases the liquid level in the riser

and liquid propagation into the pipeline. Gradually, as the pressure in the pipeline

increases due to the gas compressibility, the gas phase pushes the liquid slug further into

the riser. The stage continues until the slug reaches the riser top (for slugs of length

greater than riser length) or until the bubble penetration stage occurs.

II. Slug production

At this stage the slug occupies whole riser length. The phase is characterized by liquid

production at relatively slow rate equal to liquid inflow from the pipe and volume

displaced by the compressed gas. Pressure distribution along the riser height remains

constant and follows approximately hydrostatic pressure distribution. The phase ends

when the liquid-gas level inside of the pipe reaches the base and the gas enters the riser.

III. Bubble penetration or gas blowout

With liquid-gas interface reaching the riser bend, the gas phase enters the riser, forming

a bubble front similar to a large Taylor bubble. The bubble, continuously expanding due

to reduced static pressure, propagates up the riser decreasing hydrostatic pressure on the

riser base, therefore involving more gas into the riser. With the gas front propagating

further into the pipe, the pipeline pressure continues declining, letting the gas expand.

Eventually, the force balance between the hydrostatic and compressed gas in broken and

the riser blowout occurs characterized by ever-increasing flowrate of the liquid into the

separator due to expanding gas flowing from the pipeline.

9

IV. Gas blowdown and fallback

The fast liquid production phase is followed by violent production of pressurized gas at

high rates, called blowdown. After the gas has expanded and the pipeline pressure drops

to its minimum, the gas velocity in the riser drops low enough to let the liquid fallback

from the riser inner surface. The liquid fallen back from the riser accumulates at the riser

bend and starts the cycle again. At certain conditions liquid fallback may not occur. In

this case the new cycle is initiated from stratified flow blocking the riser bend cross

section.

The description somehow varies between the researchers, attributing the bubble penetration to

the fourth stage or distinguishing a fifth one. However, the principle remains the same.

Based on the description given above, one can infer some of the necessary preconditions for

severe slugging. First, severe slugging is sensitive to pipeline-riser geometry, with the pipeline

declination being an important precondition. Otherwise, the gas would be redistributed and

moved further down the stream due to the lower density compared to the liquid phase; therefore,

the condition under which the gas balances out and allows the liquid build up in the riser would

not be met. In addition to that, stratified flow regime in the pipeline section preceding the riser

must be present. As was shown by Schmidt [6], normal slug flow in the pipeline and riser

generated slug flow are mutually exclusive since pipeline slugs and bubbles would pass the riser

base almost unchanged, eliminating possibility for liquid accumulation at the riser base.

Second, for the riser-generated slugging to be strictly classified as severe slugging, the rate of

increase of hydrostatic pressure must be greater than the rate of increase in pipeline pressure to

the point when the liquid column reaches the riser outlet. If the condition is not met, the flow

must be considered as being a transitional to severe slugging [3]. However, transitional slugging

may still present enough hazard even with lower fluctuations in pressures and production rates

compared to pure severe slugging.

Third, the accumulated liquid column in the riser must be stable. It means that initiation of the

bubble penetration phase won’t cause rapid expansion of the gas phase resulting in blowout.

The stability criteria is met when increase of gas flowrate at a given liquid flowrate through the

vertical pipe causes increase of the pressure drop over the pipe. In other words, the gas does not

produce gas-lift effect.

10

Linga [7] pointed out at an alternative mechanism for severe slugging generation related to slugs

formed upstream of the pipeline-riser junction. Thus, a train of terrain-induced slug of sufficient

length decelerated at the riser base was observed to be a possible reason for severe slugging

initiation. The slug stopped near the riser blocks the riser base cross section and initiates the

process. Slugs arriving afterwards contribute to build up of the liquid column inside the riser.

Eventually the liquid column, being in near unstable state, is destabilized by an additional slug.

The gas entrapped between the column and the fresh slug forms a buffer, pushing the column

into the riser and initiating blowout, and decelerating the slug at the junction point. Thus, the

process repeats itself.

Possibility of severe slugging without stratified flow being present in the pipeline could

seriously affect the phenomenon prediction since slug flow occurs for different combination of

superficial gas and liquid velocities. Moreover, some of the elimination methods discussed

further in the thesis could prove to be not feasible or less effective, e.g. mixing and self-lift.

2.2. Severe slugging classification

The early works on the topic defined severe slugging as a riser induced slug flow resulting in

slugs of length equal or greater than the riser length. Those flow regimes with gas blowdown

before the liquid level had reached the riser top, therefore had slug length less than that of the

riser, were classified as transitional to severe slugging flows.

However, even transitional to severe slugging flow can be “severe” enough to call for actions.

A riser induced slugging classification by Tin and Sarshar [8], introduced in 1993, is often used

as a standard. It is presented in the following paragraphs.

Severe slugging of type 1 (SS1) is the regime we referred to before as severe slugging. During

SS1 the slug reaches the riser top prior to blowout. This type of severe slugging has a distinctive

feature of constant liquid production phase with approximately constant rate, during which the

riser base pressure remains roughly equal to hydrostatic head of the riser.

Severe slugging of type 2 (SS2) is characterized by blowout occurring prior to the liquid slug

reaching the riser top and was referred to before as transitional to severe slugging.

11

Severe slugging of type 3 (SS3) is initiated by transitional slugs generated in the upstream

pipeline and traveling along the riser, partially falling down and forming highly aerated liquid

slug at the riser base. The slug then pushed into the riser by increasing backpressure from the

pipeline side. Eventually, the main gas cap (in contrast to smaller gas bubbles, which keep

passing to the riser throughout the whole process) penetrates into the riser, starts rapid

expansion, and pushes the aerated liquid slug up the riser. Blowdown follows, generating new

portion of transient hydrodynamic slugs inside of the pipeline and the cycle repeats. In the

previous section we described this type of slugging as Linga’s mechanism.

2.3. Mass transfer effect on severe slugging

So far, the discussion on the topic only considered systems containing fluids which do not allow

mass transfer between them. The majority of studies published to date restrict their effort

exclusively on that type of system. It may be explained by two factors. First, the phenomenon

may be easier described and modeled by neglecting the mass transfer term, however at the cost

of the model accuracy. Second, such models are easier to test and verify in laboratory conditions

due to the fact that mass transfer primarily occurs due to pressure changes. Therefore, a test

setup would have to be pressurized that greatly increases the cost of experimental campaign.

Thus, the majority of the publications on the topic use air-water (and some other liquid-gas

combinations) flow loops to verify their models.

However, there is a number of difficulties related to extrapolation of such results onto real

production system. Those primarily arise due to the fact that the real production system have

pipeline lengths of the order of kilometers, compared to a fraction of that in test loops [9]. Even

more important is the effect imposed by significant pressure drop due to friction losses and

especially hydrostatic pressure drop. Pressure in turn affects physical properties of the fluids

and solvability of the gas fraction in oil.

Severe slugging in real production systems, being primarily governed by gravitational forces

and pipeline gas compressibility, exhibits high pressure variations in the riser with potential

release of the solved gas; therefore, mass transfer effect must not be ignored in modeling of the

phenomenon in subsea production systems [9].

12

A recent paper by Nemoto and Balino [9] provides a mathematical model for severe slugging

description with mass transfer between oil and gas based on Blackoil approximation for

relatively light oils (𝐴𝑃𝐼 < 45) at isothermal conditions. Simulations based on the model

showed the same evolution of the phenomenon through the four severe slugging stages;

however, significant changes in the behavior compared to air-water system were observed [9].

13

The model assumed three-phase flow consisting of oil, water and released gas, where oil and

water have the same absolute velocity and exist in homogenized flow. Mass transfer between

oil of API equal to 19 and gas of specific gravity equal to 0.6602 was allowed and calculated as

Figure 2.2. Severe slugging of type 1 cycle with effect of mass transfer[9].

A thorough description is provided in text.

14

an equilibrium state at given PVT conditions according to Blackoil model with initial GOR of

145 [9]. The production system was modelled as 1 km pipeline with 2 degree declination

followed by catenary riser of total height of 1300 m; both had internal diameter of 4 inches.

Some qualitative differences were observed over the course of one severe slugging cycle. The

following description of one cycle of SS1 is based on [9] and refers to Figure 2.2.

At the initial conditions, chosen to be when the minimum riser base pressure is observed, the

riser is partly filled with highly aerated liquid due to partial gas penetration over the riser bend

followed by expansion and accompanied by vaporization of some of the gas from the oil, see

Figure 2.2.(a). The upper level of the liquid accumulation starts to move down due to relative

motion of the gas bubbles and decrease of gas volumes penetrating into the riser, whereas the

lower level remains at the riser bend point (b). Then the liquid slug starts moving into the

pipeline, completely blocking gas passage (c). At some point all the gas phase has moved to the

top of the liquid slug or has condensate into the liquid; no gas observed over the slug length (d).

The holdup continues to remain equal to unity until the slug reaches the riser top. As the slug

reaches the riser outlet, some gas evaporates from the liquid (e-f). Liquid production phase

continues until the point when the pipeline gas reaches the riser bend and initiates blowout

(g).The gas front moves up the riser, depressurizing conduit (h). Blowdown occurs when the

gas is at the riser top, characterized by high gas flowrate over the stratified flow in the pipe and

annular flow in the riser, inducing new portion of transient slugs in the pipeline and initiating

the cycle again (i).

An important consequence of the described process is that the liquid production is always

associated with some of that of gas due to mass transfer from liquid to gas, opposite to no-mass-

transfer case when pure liquid production is possible [9]. It also affects the time profile of riser

base pressure, see Figure 2.2.(j). Despite the fact that the process may be seen as SS1, there is

no plateau of constant pressure during liquid production phase. That in turn is important for

understanding the process as well as flow regime identification in the following parts of the

thesis.

15

2.4. Severe slugging occurrence criteria

Steady state models consider the balance of forces acting on the fluids in pipeline-riser system

in conditions close to severe slugging. The main purpose of the steady state models is to find

the criteria necessary for severe slugging occurrence. They usually neglect friction forces due

to stationary conditions, and assume constant inflow rates.

It is worth mentioning prior to discussion that due to steady state nature, the models consider

only severe slugging resulting from the classical severe slugging mechanism and incapable of

describing the system behavior with fast variations pipeline holdups, such as that stated by Linga

[7].

Under the stationary assumption, three main conditions must exist for severe slugging to form

[3]:

1. stratified flow in the negatively inclined pipeline section preceding the riser base;

2. rate of hydrostatic pressure increase in the riser due to liquid build up must be higher

than the rate of pipeline gas pressure increase;

3. the liquid column instability inside the riser.

Each of the criteria below address to one of the three preconditions.

2.4.1. Stratified flow criterion

Classical mechanism for severe slugging occurrence requires stratified flow to be present in the

pipeline. The criterion for transition to stratified flow in relation to severe slugging was

described by Taitel and Dukler [10]. They considered conditions necessary for wave generation

in two phase flow between parallel plates and extended the model to the flow in circular pipe,

taking into account interaction between the generated waves and gas flow accelerated over the

wave crests. The criterion, expressed in terms of superficial gas velocity, can be written as [10]:

𝑈𝑆𝐺 < 𝐾2 [(𝜌𝐿 − 𝜌𝐺)𝑔 cos 𝛽 𝛼𝐺

𝜌𝐺 𝑑𝛼𝐿/𝑑ℎ𝐿]

0.5

, (2.1)

where 𝐾2 ≈ 𝛼𝐺/𝛼𝐿 ,

𝑑𝛼𝐿/𝑑ℎ𝐿 – change of liquid flow area with liquid height.

16

When the criterion is followed, stratified flow is expected to be observed in the pipeline.

The relation expressing the liquid holdup with respect to liquid height may be derived based on

geometrical considerations [10]:

𝑑𝛼𝐿𝑑ℎ𝐿

= 𝐷√1 − (2ℎ𝐿𝐷− 1)

0.5

. (2.2)

2.4.2. Bøe criterion

The Bøe criterion [11] focuses on the second precondition for severe slugging, namely that the

rate of increase of hydrostatic pressure in the riser must be greater than the rate of increase of

the pipeline pressure. The criterion, assuming constant gas and liquid flowrates, can be written

as [11]:

𝑈𝑆𝐿 ≥𝑃𝑃

𝜌𝐿𝑔(1 − 𝛼𝐿)𝐿𝑈𝑆𝐺 . (2.3)

To resolve the inequality, the liquid holdup must be evaluated from the superficial gas and liquid

velocities. The first approximation by Bøe [11] assumed no-slip condition between the phases,

resulting in identical absolute speeds and equation for the liquid holdup in the form of:

𝛼𝐿 =𝑈𝑆𝐿

𝑈𝑆𝐿 + 𝑈𝑆𝐺. (2.4)

This condition results in a straight line on the superficial velocities graph, distinguishing severe

slugging and no severe slugging regions.

More rigorous correlations for liquid holdup, such as that presented in Taitel [12], result in an

operational envelope outlining the severe slugging region. Complete derivation of the iterative

steps for finding the correlation may be found in the Appendix of Montgomery [13].

2.4.3. Pots criterion

Pots et al. [14] formulated a criterion for liquid build up in the riser resulting in severe slugging.

Similarly to the Bøe’s criterion, it considered the pressure balance between the hydrostatic

pressure of the liquid in the riser and accumulated in the pipeline gas pressure, under assumption

that all of the liquid entering the flow line goes to the riser. The resulting condition may be

written as:

17

П𝑆𝑆 =𝑧𝑅𝑇/𝑀

𝑔𝛼𝐿𝐿

�̇�𝐺

�̇�𝐿. (2.5)

Severe slugging of type 1 occurs when П𝑆𝑆 < 1. Additionally, Pots et al. [14] proposed to use

the measure of П𝑆𝑆 to determine the degree of slugging severity, where smaller values

correspond to more harsh slugging conditions.

Since both Bøe and Pots criteria are based on the same principle, it may be shown that the Pots

criterion takes form of the first one when expressed in terms of pressure and superficial

velocities [13].

2.4.4. Taitel criterion

Taitel [12] speculated on the stability of operations under conditions that are close to blowout

(precondition 3 discussed above). He considered stability of the liquid column in the riser at the

moment of first bubble penetrating the riser bend. If the column is not stable, the bubble will

cause the full cycle of blowout, sweeping almost all the liquid from the riser. Otherwise, a

bubble or slug flow will occur.

To derive the criterion, the difference of the forces acting on the column before and after the

penetrating bubble front was considered. The force difference may be written as [12]:

∆𝐹 = [(𝑃𝑠𝑒𝑝 + 𝜌𝐿𝑔ℎ𝑅)𝛼𝐺𝑃𝐿

𝛼𝐺𝑃𝐿 + 𝛼′𝐺𝑦] − [𝑃𝑆 + 𝜌𝐿𝑔(ℎ𝑅 − 𝑦)], (2.6)

where 𝛼′𝐺 stands for the gas holdup in the gas cap penetrating the liquid column.

Then the criterion for stability becomes [12]:

𝜕(∆𝐹)

𝜕𝑦< 0 (2.7)

for 𝑦 = 0 (at the riser base).

Combining the two equations above and referencing to atmospheric conditions (to express gas

flow rate in terms of superficial gas velocity at standard conditions) the criterion takes form

[12]:

𝑃𝑠𝑒𝑝

𝑃0>(𝛼𝐺𝑃/𝛼′𝐺)𝐿 − ℎ𝑅

𝑃0/𝜌𝐿𝑔. (2.8)

18

The resulting criterion is mainly dependent on the pipeline-riser system geometrical parameters

and the separator backpressure. The gas holdup values vary within 0.8 to 1.0 interval for

stratified flow [12]. The model is relatively insensitive to the value of 𝛼′𝐺, and Taitel [12]

advocates that 0.89 is a good assumption. The pipeline gas holdup, 𝛼𝐺𝑃, may be expressed in

terms of superficial velocities under the stratified flow model, allowing to plot the criterion on

a flow regime map.

2.5. Effect on production system

Slug flow in general may impose significant problems on the production system. As a rule of

thumb, the longer the slugs the greater negative effects imposed. Therefore, severe slugging

may be considered as the most violent form of slug flow. The three parameters of main concern

associated with riser induces slug flow are pressure fluctuation, high superficial velocities, and

variability in production rate.

High variability of pipeline pressure affects the field productivity, imposing additional

backpressure on the producing wells. In general, resulting loss in production is system specific

and may vary greatly. In some extreme cases severe slugging may kill a well or cause putting it

on hold [15]. Additionally, it causes high stress loads on riser pipework and topside piping [16].

Variability of the production rates on the platform is, by far, the main challenge of severe

slugging. As it was demonstrated above, severe slugging is characterized by periods of no

production, slow liquid production, and increasing liquid production, reaching extreme flow

rates and followed by violent gas blowdown. Pots et al. [14] states that the liquid production at

its peak may reach velocities as high as 70 times its average value. As the result, increased

vibration, erosion, and general wear and tear rates are experienced by the system. It is especially

true for erosion-corrosion processes since they are strongly dependent on the media flow speed.

Moreover, droplet induced erosion may be caused by high gas flow rate even without solids

being present in the flow [17].

At the same time, variability in the production rate imposes large disturbances in the separator

train, decreasing the separator effectiveness, resulting in poor separation of the phases, water

carryover to the export system, and improper water treatment, leading to oil content in

discharged water and potential environmental requirements violations [15]. Extreme cases may

19

result in separator flooding, emergency shutdowns, and liquid carryover to the process flare

system [16]. Moreover, gas flow rate instability causes inefficient compressor operation,

resulting in increased maintenance costs and higher flare gas volumes [15].

20

3. Severe slugging elimination

3.1. Conventional methods

3.1.1. Backpressure increase

As it may be inferred from Bøe and Pots criteria, severe slugging of type 1 may be alleviated by

increasing the backpressure imposed on the system, thus increasing the “stiffness” of the gas

phase resulting in earlier initiation of the blowout phase of severe slugging cycle. However, the

method cannot stabilize the flow completely, merely reducing the length of the slugs produced.

In addition to that, the method is connected to sufficient losses in production volumes. Yocum

[18] reports flow capacity reduction up to 70% in some cases. For that reason the technique was

deemed to be inefficient even for low depth offshore developments [19].

3.1.2. Choking

Choking as a way to control slug flow in vertical conduits was considered relatively early in the

history of multiphase flow production. Thus, Yocum [18] in 1973 elaborates on choking

efficiency for slug flow mitigation and states that at that time the method was used inefficiently,

inducing excessive pressure losses and decreased production rates.

Topside choking addresses to the flow stability criterion described in Section 2.4.4.

Combination of choke and flowline characteristic curves allows reduction of the minimum

required gas flowrate to stabilize the flow at the cost of higher pressure drop, see Figure 3.1.

21

Schmidt [20] recognized topside choking as a means for severe slugging alleviation and later

[6] provided a description of the control mechanism. Proportional to the liquid velocity increase

of backpressure due to a choke installed topside creates an additional retarding force on the gas

phase during the bubble penetration phase. Suppressed bubble could not accelerate to the level

of blowdown initiating. Therefore, normal slug flow regime was forced instead of severe

slugging.

A more thorough description of the choking mechanism with theoretical and experimental

investigation was done by Jansen [21-24]. The work was based on a modified Taitel’s stability

criterion [12], taking into account additional effect of the choke. Thus, backpressure upstream

from the choke right before the bubble penetration occurs may be written as:

𝑃𝐵 = 𝑃𝑠𝑒𝑝 + 𝐶𝑈𝑆𝐿2 , (3.1)

where 𝐶 – choke coefficient, [Pa∙s2∙m-2].

An important assumption of the model for two-phase flow is that time averaged pressure drop

over the choke is dependent only on liquid phase and superficial liquid velocity. Therefore, the

pressure drop as a function of time is assumed to be a function of both liquid holdup and the

Figure 3.1. Flow stabilization by topside choke introduction [6].

The graph shows the result of characteristic curves combination plotted at constant

superficial liquid velocity. It may be seen that superficial gas velocity guaranteeing stable

flow is reduced by a factor of 10.

22

mixture velocity [21]. These statements was reported to be verified by experiments of Jansen

[21]; however, no specific data about the experiments were provided.

When the gas phase is entering the riser, it creates an additional response from the choke due to

acceleration of the liquid column, which can be proved to be linearly dependent on the gas

penetration height [21]:

𝑃𝐵 = 𝑃𝑠𝑒𝑝 + 𝐶𝑈𝑆𝐿2 + 𝐾𝑦, (3.2)

where 𝐾 – proportionality constant, [Pa∙m-1].

Similar to the stability criterion discussed in Section 2.4.4, the force per area difference applied

to the interface between the penetrating gas phase and the bottom end of the liquid column may

be written as [21]:

∆𝐹 = [(𝑃𝑠𝑒𝑝 + 𝐶𝑈𝑆𝐿2 + 𝜌𝐿𝑔ℎ𝑅)

𝛼𝐺𝑃𝐿

𝛼𝐺𝑃𝐿 + 𝛼′𝐺𝑦]

− [𝑃𝑠𝑒𝑝 + 𝐶𝑈𝑆𝐿2 + 𝐾𝑦 + 𝜌𝐿𝑔(ℎ𝑅 − 𝑦)].

(3.3)

Following the same logic as it was presented in Section 2.4.4, stability criterion may be found

by differentiating the equation above, bearing [21]:

𝑃𝑠𝑒𝑝 + 𝐶𝑈𝑆𝐿2

𝑃0>

𝛼𝐺𝑃𝛼′𝐺

𝐿(1 −𝐾𝜌𝐿𝑔

) − ℎ𝑅

𝑃0/𝜌𝐿𝑔. (3.4)

Relation between the choke coefficient and the proportionality coefficient may be found from

the equation of motion written for the liquid column in the riser with respect to applied force

∆𝐹 [21]. The solution brings [21]:

𝐾 =2𝐶𝑈𝑆𝐿

2

ℎ𝑅. (3.5)

The model, however, only describes flow stability at the initial point when the riser is fully

occupied by the liquid phase. When the steady state operations is established, some fraction of

the gas will be present in the riser. Therefore, the average density over the riser is reduced.

Neglecting the gas density compared to that of liquid, the average density may be approximated

to 𝜙𝑅𝜌𝐿. The stability criterion then modified as follows [21]:

23

𝑃𝑠𝑒𝑝 + 𝐶𝑈𝑆𝐿2

𝑃0>

𝛼𝐺𝑃𝛼′𝐺

𝐿(𝜙𝑅 −𝐾𝜌𝐿𝑔

) − 𝜙𝑅ℎ𝑅

𝑃0/𝜌𝐿𝑔. (3.6)

Introduction of a choke into the system increases the flow stability, contracting the unstable

flow operational envelope along the superficial liquid velocity axis: the greater the choke

coefficient 𝐶, the smaller the value of the transitional velocity from unstable to stable flow.

Figure 3.2 shows how different choke valve settings affect the flow regime.

Figure 3.2. Choking effect on the severe slugging envelope [21].

Fig. a) shows stability region for choke set to C=120000 Pa∙s2∙m-2 and b) for C=245000

Pa∙s2∙m-2. It may be seem how with increased choke resistance, severe slugging envelope

contracts along liquid velocity axis.

a)

b)

24

3.1.3. Riser-base gas injection

The method of gas injection for severe slugging remediation is based on the artificial gas lift

principle and, according to Mokhatab and Towler [25], was one of the most frequently used.

Gas injected in the riser base is deemed to lower the static pressure of the liquid column and

shift the flow regime to annular or dispersed flow, thus solving the problem of slugging.

However, Mokhatab and Towler [25] highlight that gas injection may be relatively useless for

transient slugging when an already formed liquid plug arrives to the riser.

Despite the fact that the method was considered not economically feasible by some researchers

at the beginning of 80s [6], some investigation was conducted. Pots et al. [14] experimentally

studied effect of gas injection on severe slugging and concluded that for injection rates of about

50% the slug arrival velocities are considerably lower than in the case without treatment.

However, about four times the amount of produced gas must have been injected to completely

eliminate slugging, approaching annular flow regime [14]. Hill [26] reported real case

implementation of gas injection on the S.E. Forties field, indicating reduction in severe slugging

extension. The method also succeeded in reviving some of the wells previously killed-off by the

back pressure from the riser slugging [27].

The method feasibility and effectiveness reduces with the water depth owing to increased

pressure loss on friction and gas compressibility. Thus, gas cooling due to Joule-Thomson effect

rising from high water depth makes the flow conditions more susceptible to wax and hydrate

formation inside of the riser [25]. In addition to that, higher water depth requires the gas injection

line to be longer, making it more costly and the friction losses more prominent.

Jansen et al. [21] described the performance of the gas lift method based on the Taitel’s stability

criterion. The modified stability criterion may be written as [21]:

𝑃𝑠𝑒𝑝

𝑃0>

𝛼𝐺𝑃𝛼′𝐺

𝐿 − ℎ𝑅

𝑃0/𝜙𝑅𝜌𝐿𝑔, (3.7)

where 𝜙𝑅 is the average liquid holdup over the riser length, that can be expressed in terms of

gas and bubble velocities:

𝜙𝑅 = 1 −𝑈𝑆𝐺,𝑅

𝑋0𝑈𝑀 + 𝑌0. (3.8)

25

Parameters 𝑋0 and 𝑌0 are the drift parameter and drift velocity, respectively; both vary

depending on the flow regime inside of the riser.

Gas injection affects the flow map diagram in a way that the severe slugging is contracted similar

to the case of choking with the difference that contraction occurs along the superficial gas

velocity axis with increasing rate of gas injection [21].

3.1.4. Active control methods

The basic principle behind all of the active control methods is to make use of online

measurement of the process and/or pipeline information to control the system by adjusting

available degrees of freedom such as choke opening level, process pressure, and levels inside of

the process facilities [15].

Feed-forward control relies on prediction of slug occurrence and preparation of the separator

train to accommodate them [15].

The method of slug flow suppression based on the active control of the choking system got a lot

of attention in 90s and was successfully introduced into practice later. The method relies on the

pressure readings either from subsea pipeline or topside facilities and based on prediction of the

slug slow occurrence. As the result, choke valve, either subsea or topside, is adjusted

accordingly and the flow is maintained within required regime. Solutions of that kind were

developed by many of the oil and gas producers and service companies, such as ABB, Statoil,

and Total [15, 28-30].

The active choking technique has a significant advantage over the classical manually operated

choking for slugging suppression since it may operate with average valve opening greater than

that of regular choking, e.g. [15]. The stability achieved due to online tuning of the valve

opening, thus maintaining desired multiphase flow regime.

Slug Suppression System (S3), belonging to Shell, was first published in 1995 [31] and assumed

installation of an additional “mini-separator” (in later modifications substituted with a piping

section [32]) prior separation of the phases with sufficient measuring of gas and liquid flowrates

and pressure, see Figure 3.3. Based on the readings and processing, separate chokes for gas and

liquid are actuated to facilitate flow stability.

26

The manufacturer states that the technology creates less backpressure on the pipeline system

compared to regular topside choking (only 1 to 3 bar pressure drop over the system in control

mode claimed [32]) and occupies considerably less space compared to a slug catcher [33].

3.2. Unconventional methods

3.2.1. Self-lift

The method of severe slugging elimination based on injection of the produced gas from the

production pipeline into the riser base, in contrast to that when the gas is treated and compressed

on the platform and then transported through a separate conduit to the riser inlet, was first

proposed by Barbuto in 1995 [34].

The statement consisted of a bypass, connecting the pipeline and the riser, allowing the flow of

gas and its injection at the predetermined position of the riser. It was said to be located at

approximately a third of the total riser height [34]. In addition, the patent describes some of the

ways to control throughput of the bypass line to alter the pipeline pressure. However, no further

information, tests or theoretical proof followed.

Figure 3.3. S3 principle scheme [33].

Based on the readings of the pressure (P) and liquid level (L) from the mini-separator

and from the gas and liquid flowmeters (F), the control system actuated the one phase

chokes.

27

An investigation of the method feasibility was conducted at the University of Tulsa and

published in the thesis of Tengesdal [35]. The experimental campaign of the thesis was run on

a test facility of 65 ft. [19.8 m] pipeline followed by 49 ft. [14.9 m] riser, both with ID of 3 in.

[76.2 mm]. Effective pipeline length for the system was equal to 280 ft. [85.3 m] due to tank at

the pipeline inlet. A conduit of 1 in. [25.4 mm] with installed choke valve was used as a bypass

allowing connection with any combination between 4 take-off points on the pipeline and 4

injection points on the riser, see Figure 3.4 [35]. Therefore, the facility allowed investigating

effectiveness of the system with variation of all the parameters of interest: gas/liquid superficial

velocities, inclination, take-off/injection point location, pressure drop over the bypass.

Overall, a stable production point was found for any combination of superficial gas and liquid

velocities and pipeline inclination with the use of choke valve installed in the bypass line.

However, a stable flow could be achieved for fully open choke valve only with relatively high

Figure 3.4. Pipeline-riser system with bypass used in experiments of [36].

The drawing shows take-off and injection points locations for bypass length

adjustments:

PCV 1-8 – pressure control valves;

BV 7-10 – ball valves to operate the take-off point location;

BV 11-14 – ball valves to operate the injection point location;

MPRV – ball valve used for choking.

28

superficial gas velocities, due to increased pressure drop over the bypass; otherwise, the system

should have been tuned with the aid of the choke valve.

It was found that situations when the injection and take-off point levels were close, stabilized

flow conditions were obtained more easily due to pressure balance between them [35]. Was the

injection point moved upwards, the take-off should has been moved upstream away from the

increased liquid penetration level into the pipeline, or, alternatively, additional pressure loss

over the bypass could be provided by the choke. Otherwise, the bypass inlet is blocked with

liquid, resulting in instable flows.

An extreme situation of another sort occurs when the injection point is too close to the riser

base. In that case, dual gas penetration may happen, one from the bypass injection point and one

from the riser base, resulting in instable flow [35].

Ironically enough, one of the necessary features for severe slugging occurrence – the pipeline

declination – has a positive effect on the self-lift severe slugging elimination technique. With

increasing level of the pipeline inclination, it is easier to achieve a stable flow. As experiments

by Tengesdal [35] show, at -1° downward angle, the stable flow was easily interrupted by

perturbations between the phases. The gas phase fingers towards the riser base and, if reaches

the riser bend, partial blowdown occurs, causing instabilities in the pressure and flow rate levels.

At higher declination angles, a better separation between the phases occurs and gas is less likely

to reach the riser via the riser base.

The self-lift technique is relatively insensitive to the liquid and gas flow rates. One of the main

limiting factors when the designing the system for different flow rates, is the distance to the

take-off point, see Figure 3.5. Presumably, it is possible to find an optimal choke and distance

to the take-off point combination satisfying all encountered flow rates during the field life.

29

3.2.1.1. Self-lift steady-state model

In his work Tengesdal [35] derived two steady-state models of the technique: a simplified and

a rigorous one. Both of them are based on the analysis of the hydrostatic balance between cross

sections A and B, see Figure 3.6.

a)

b)

Figure 3.5. Flow rate sensitivity with constant superficial liquid (a) and gas (b) velocities

[35].

Declination level -3°. Injection point: BV13; take-off point: BV10.

30

The simplified model assumes no pressure loss on friction and, constant PVT and flow

properties of the fluids due to small variations in pressure and temperature within area of

interest, and that density of the gas, which is much less then density of liquid, may be assumed

to be zero [35].

The resulting criterion for stable flow with installed bypass may be written as [35]:

𝑑𝑃𝑠𝑒𝑝

𝑑𝑡+𝑔

𝐴𝑅�̇�𝐿 ≤

𝑧𝑅𝑇

𝑀(𝑙𝑃 − 𝑙𝐵𝑦𝑝𝑎𝑠𝑠)𝐴𝑃𝛼𝐺,𝑃�̇�𝐺 . (3.9)

The criterion is similar to that of Bøe for severe slugging occurrence in the riser system.

The more strict criteria takes into account the flow pressure loss over the distance and the riser

base elbow. The new limiting criterion becomes the bubble penetration length, 𝐿𝐵𝑢𝑏, which

must lay within 0 ≤ 𝐿𝐵𝑢𝑏 ≤ 𝐿𝐵𝑦𝑝𝑎𝑠𝑠. If the bubble length is less than 0, meaning that the bubble

recedes to the point above the take-off point, and the liquid enters the bypass. The bubble length

of greater than 𝐿𝐵𝑦𝑝𝑎𝑠𝑠 would mean that the gas front reaches the riser base elbow [35]. Both of

the scenarios would result in unsteady flow with slugging in the riser of sufficient length.

Figure 3.6. Schematic depiction of the self-lift system [36].

31

Therefore, to satisfy the steady flow criterion, the following system of equations must be

resolved with respect to the parameters of interest for design process [35]:

{

∆𝑃𝐿,𝑚𝑖𝑛 ≤ ∆𝑃𝑏𝑦𝑝𝑎𝑠𝑠 ≤ ∆𝑃𝐿,𝑚𝑎𝑥

∆𝑃𝐿,𝑚𝑎𝑥 = 𝜌𝐿𝑔(𝑦 − 𝐷𝑃) +2𝑓𝐿,𝑅𝜌𝐿𝐷𝑅

(𝑄0𝐴𝑅)2

(𝑦 − 𝐷𝑃) + ∆𝑝𝑒𝑙𝑏𝑜𝑤

∆𝑃𝐿,𝑚𝑖𝑛 = 𝜌𝐿𝑔(𝑦 − 𝐷𝑃 − 𝐿𝑏𝑦𝑝𝑎𝑠𝑠𝑠𝑖𝑛𝜃) +2𝑓𝐿,𝑃𝜌𝐿𝐷𝑃

(𝑄0𝐴𝑃)2

𝐿𝐵𝑢𝑏,𝑃 +

+2𝑓𝐿,𝑅𝜌𝐿𝐷𝑅

(𝑄0𝐴𝑅)2

(𝑦 − 𝐷𝑃) + ∆𝑃𝑒𝑙𝑏𝑜𝑤

∆𝑃𝑏𝑦𝑝𝑎𝑠𝑠 = ∆𝑃𝑏𝑦𝑝𝑎𝑠𝑠,𝑓𝑟𝑖𝑐 + ∆𝑃𝑏𝑦𝑝𝑎𝑠𝑠,𝑐ℎ𝑜𝑘𝑒

(3.10)

According to Tengesdal [35], the model predicted accurately occurrence of severe slugging. The

accuracy of the model was low in cases when the liquid-gas interface was not stable enough.

When liquid penetration level is close to either the take-off point or the riser base, oscillations

of the interface could cause entering of the liquid into the bypass or partial blowdown,

respectively. The effect was less prominent with increased pipeline declination, as it was

mentioned earlier. Thus, as Tengesdal [35] demonstrates, for declination of -1°, 25 out of 92 fell

outside of the prediction model; for declination -3°, 5 out of 151 tests were predicted erroneously

by the model; for declination of -5°, all 162 experimental points were predicted correctly. The

outliners corresponded to the cases when the liquid-gas interface instability could affect the

stable operation of the system.

3.2.2. A conduit of reduced diameter

The method relies on the principle that the gas and liquid superficial velocities may be increased

by reduction of the conduit diameter, moving the operational point out of severe slugging

envelope. As early as 1973, Yocum [18] proposed to make use of this fact by using a riser with

smaller diameter or substituting it with a number of such risers.

Wyllie and Brackenridge [37] proposed to use an insertion of a smaller diameter pipe inside of

the riser for severe slugging mitigation. The annulus (number 2 on Figure 3.7) was presumably

used as a gas injection point further enhancing effectiveness of the concept. However, the

insertion could impose restriction on pigging, which is an important flow assurance measure for

wax deposits cleaning, especially for deep waters [19]. Later a patent was filled by Wyllie [16]

32

for an apparatus of such a design. The patented solution allowed the insertion to be retrievable,

however a workover operation was needed. The conceptual drawing is presented on Figure 3.7.

3.2.3. Flow conditioning, phase agitation, mixing

The concept of remixing of the multiphase flow right before the riser entrance for slug flow

alleviation was first mentioned by Yocum [18]. He proposed to use either helices with properly

selected pitch, turbulence inducing flow restrictors (bafflers or impact mixing barrels), or

agitation by mixers rotated from the platform [18].

The purpose of a mixing tool being present in the conduit is to break up the stratified flow inside

of the pipeline, which is one of the preconditions for severe slugging. Entering the riser in form

of droplets, liquid, assuming that the gas velocity inside of the riser exceeds the critical velocity,

cannot accumulate at the riser base, creating the conduit cross section blockage, therefore, the

classical mechanism for severe slugging initiation is eliminated.

Figure 3.7. Smaller diameter conduit insertion [16].

The riser insertion device. Perforations denoted on the figure by number 8 allow gas

injection into the flow.

33

The fact that the technique is passive makes it attractive for implementation in extremely remote

conditions, such as for terrain induced slugging mitigation in horizontally drilled wells, see [38].

Recently, a series of studies on that topic was conducted by Brasjen et al. [39, 40]. Number of

passive devices, such as mixers, swirls, perforated liners and choke, was investigated. The

researchers highlight quick restructuration of the mixed flow downstream the mixing devices.

Therefore, it was concluded that the optimal placement of such a device is near to flowline

undulation dip. Reduction of slugging frequency up to 16% in large scale test facility was

observed, however, at the price of significant pressure drop over the system [39].

3.2.3.1. Phase agitators

Some of the proposed designs assumed possibility of the multiphase flow homogenization by

introducing some kind of obstacle to the flow stream, by that facilitating the mixing processes

in the stream, as it was mentioned referring to Yocum [18].

To that category the patent after Arnaudeau and Corteville [41] may be attributed. The authors

proposed a train of mixing members of a defined shape to be placed inside of the multiphase

conduit. The members were to be jointed on a rod and, supposedly, lowered down by means of

a wireline, see Figure 3.8.

However, no further information on the topic may be found in literature. In 1992 the patent was

ceased due to non-payment of the annual fee, probably indicating abandonment of the invention.

Figure 3.8. Mixing members inside of a conduit [41].

The three circular cross sections give an example of mixing members geometry.

34

3.2.3.2. Undulation

Despite the fact that undulating shape of the pipeline may be one of the preconditions to severe

slugging [21, 38], similar to the effect of downwards inclination of the pipeline, some research

interest was generated around the altering of the pipeline shape as a passive method for severe

slugging alleviation.

The patent by Makogon and Brook [42] proposes installation of at least one section of pipeline

configuration, consisting of a positive inclined pipe followed by a horizontal and declining

section, see examples on Figure 3.9. It was claimed that the setup of proper configuration and

placement allows breaking up of severe slugging flow into smaller slugs, at the same time not

restricting pigging operations due to appropriate bending radiuses [42].

The device was tested by Makogon et al. [43] over the commercial transient multiphase flow

simulator OLGA and on scaled down model at the University of Tulsa. The concept proved to

be viable and provided severe slugging alleviation, reducing the pressure fluctuations over the

riser due to generation of smaller slugs.

Another undulating pipeline design is mentioned in the patent belonging to Shen and Yeung

[44], which describes a pipeline configurations to serve the purpose of stratified flow mixing,

assembled from consecutively jointed tubular segments with predetermined angles, forming the

shape of a spiral or a wave, see Figure 3.10.

Figure 3.9. Examples of flow conditioning by Makogon and Brook [42].

35

It may be inferred from the patent [44] that the shape depicted in the middle of Figure 3.10 (later

called “wavy pipes”) showed the best results over others, providing the greatest reduction in the

operational region attributed to severe slugging.

Additional research was conducted by Xing et al. [45-47] on the effectiveness of the wavy pipes

for slugging alleviation. The wavy section was assumed to be made of standard piping sections,

therefore allowing pigging operations and having lower manufacturing costs.

Experimental study on the topic published in 2013 by Xing et al. [47] proved the concept to be

relatively efficient in the small scale model. The authors claim that the presence of the wavy

pipe section before the riser base consistently reduce the severity of slugging by reducing the

slug length and the severe slugging occurrence region on the flow regime map. For a number of

Figure 3.10. Undulating pipes design proposed by Shen and Yeung [44].

Each section of the pipe is a standard piping bend of 90 degree.

36

test trials, transition from severe slugging to oscillatory flow was achieved. It was also found

that the pressure drop over the system with wavy pipes installed in it was consistently lower

compared to the case without those. The number of the up-down section was showed to have

correlation with the method efficiency, provided some distance between the wavy section outlet

and the riser base.

3.2.3.3. Venturi tube

An interesting approach for severe slugging elimination was proposed by Almeida and Goncalez

in 1999 [48]. The patent assumed installation of a venturi tube near the riser base for temporary

rearrangement of stratified flow into non-stratified flow over the distance sufficient to overpass

the riser bend.

The device represents a venturi tube consisting of convergent-divergent nozzle installed inside

of the conduit, as it is shown on Figure 3.11.

In addition to stratified flow rearrangement the patent claims that the device, due to sudden drop

in pressure in the throat part of the nozzle, is capable to release significant quantities of the gas

dissolved in the liquid phase [48]. According to the claim, it would further contribute to

establishment of a non-stratified flow regime.

The consecutive paper [49] on the experimental investigation of severe slugging elimination

effectiveness with venturi demonstrated that pressure fluctuations may be suppressed with the

device, provided that it has sufficiently small throat diameter. The test trials comparing the

venturi to the choking showed almost identical results.

Figure 3.11. Venturi tube for stratified flow [48].

37

Criticism.

The paper claims that venturi would impose “the smallest pressure drop through the device”,

despite the fact that only experiments on water and air were conducted [49]. For the real case

scenario this statement is rather doubtful. Such effects as mass transfer between the phases and

possible presence of water in oil-gas flow, forming highly viscous dispersion with oil, must be

taken into account to make a sound statement about the pressure loss over a venturi tube in

multiphase flow.

Second, the paper [49] compares effectiveness of choking to venturi in severe slugging

mitigation. However, no information is given considering the choke setup; therefore, no

judgment may be given about how efficiently the choke was set.

Third, the paper [49] considers devices with two different throat diameters: 6 and 8 mm. The

former one eliminated severe slugging, whereas the latter failed to do so. To illustrate the effect,

authors plotted three points on the flow pattern map, representing regimes with the two devices

(6 and 8 mm throats) and without it. The points plotted correspond to the superficial velocities

inside of the throat (it may be inferred from the fact that all three points are plotted along a line,

having the angle of 45 degrees and from the statement given in the paper and from the wording

of the paper). It is highly doubtful that such an assumption is valid for description of the overall

regime in the pipe. Superficial velocities will return to their initial values after they exit diffusor,

especially taking into account “the smallest pressure drop through the device” (no sufficient gas

compression happens; the statement is always true for incompressible fluid, given constant

boundary conditions of the conduit system).

Based on the arguments presented above, the approach chosen by the authors to describe the

proposed may be questioned.

3.2.4. Separation

Presence of at least two phases with different densities is necessary for severe slugging

occurrence. Therefore, terrain slugging, in principle, must be completely eliminated by

separation of the liquid and gaseous phases. However, some researchers point out at observed

intermittent flow regimes in pipelines transporting only oil and water, due to the fact that oil is

more viscous and less dense and the two liquids are immiscible.

38

The concept of subsea separation may eliminate slugging flow in the riser and increase

production; however, it is associated with a number of problems. First, after the gas phase is

separated in must be transported via an additional conduit. Provided that there is no pipeline

or/and riser available for this purpose, the concept may turn out to be economically not feasible.

Benetti and Villa [50] propose to utilize an existing pipeline of two parallel pipelines from a

field, one of which was abandoned due to flow assurance considerations, when a subsea

separation system is installed near to wellhead.

In addition to that, the liquid flow separated from gas will impose additional backpressure due

to increased averaged density in the riser, therefore increased hydrostatic pressure. For that

reason a typical subsea separation system must include a pump for the liquid phase boosting.

Operations of a pump added to that of a separator may be considered troublesome.

There are a number of benefits associated with subsea separation and boosting apart from

slugging elimination and production increase. Thus, implementation of the technique may be

economically efficient due to reduced investments in flow assurance management. Two-phase

separation, boosting, and further transportation via separate flowlines depressurizes the system,

considerably reduces amount of water in gas-phase flowline, and significantly decreases Joule-

Thomson cooling of the entrained liquid, all of which positively affect hydrate curve, lessening

the required amount of glycol injected [27]. Potential benefits from subsea separation rise with

increase of water depth and proved to be important decision drivers for such projects as Perdido

[51], Parque das Conchas [52] and Pazflor [53].

Three-phase separation (with additional separation between oil and water) brings additional

benefits such as potential reduction of topside capacity and backpressure and possibility to re-

inject the separated water in situ. However, it goes beyond the scope of the thesis.

The remaining portion of the section gives a brief overview of two-phase subsea separation and

boosting solutions that have already been tested and proved their effectiveness. For information

on other subsea separation technologies not involving the boosting component consult

Hannisdal et al. [27].

39

3.2.4.1. Vertical annular separation and pumping system (VASPS)

Vertical Annular Separation and Pumping System (VASPS) concept was proposed at the

beginning of 90s as a subsea solution combining two phase separation and liquid phase boosting.

The system had significant height and was supposed to be installed and secured in vertical

position inside of a 30 or 36 in. conductor pipe in a dummy well of approximate depth of 30-70

meters, drilled and completed by conventional techniques [50].

The VASPS body unit consisted of three concentric casings [50]:

1) Outer casing with nominal diameter of the conductor pipe provided pressure integrity of

the system;

2) Helix Separator Joint (intermediate casing) consisted of a 16-20 in. pipe with

perforations and welded to the outer surface helix plates, spiraling to the bottom of the

assembly and being in sufficient sealing contact with the inner surface of the outer

casing;

3) Liquid Discharge Tubing (inner conduit) was a 8-10 in. pressurized conduit for the liquid

phase pumped from the bottom of the assembly by an ESP, accommodated inside the

tubing.

Thus, a liquid-gas mixture enters the outer-intermediate casing annulus and forced to flow in

spiral due to gravity drainage and initial momentum. Under the action of gravity and induced

centrifugal forces the gaseous phase is separated and escapes the helix through the perforations

in the intermediate casing, entering the annulus between the intermediate casing and inner

tubing from which it is directed to the gas pipeline. The liquid phase is drained to the bottom of

the assembly where it completes its degassing in a sump and then pumped by the ESP into the

inner tubing and directed into the liquid phase pipeline to the process facilities. Figure 3.12

provides schematic illustration of the VASPS.

40

Initial phase of the field test of the system in 2001 at the Marimba field, Brazil, proved VASPS

to be a feasible solution. Installation of the system on the well MA-01 reduced the wellhead

pressure from 36 to 11 bars, increasing production from 750 m3/d to 1000 m3/d and eliminating

the need for injection of 100,000 m3/d gas-lift gas [55].

3.2.4.2. Caisson-ESP (C-ESP)

Caisson-ESP concept is similar to that of VASPS. It consists of a vertical caisson of

approximately 100 meters of height secured on the seabed inside of a dummy well drilled and

completed for that purpose. A multiphase flow from producing wells enters the assembly at the

top through a specifically designed angled and tangential inlet. Thus the fluid enters cylindrical

cyclonic separator where the gas is separated from the liquid phase under the action of

gravitational and centrifugal forces. The liquid falls to the bottom of the caisson where it is

pumped by an ESP into a dedicated liquid phase outlet, whereas the gaseous phase expands into

the gas outlet [52], see Figure 3.13.

Figure 3.12. Schematic depiction of the VASPS principle and main components [54].

41

The concept was implemented on Perdido [51] and Parque das Conchas [52] fields. Iyer et al.

[52] report that implementation of C-ESP on Parque das Conchas allowed efficient severe

slugging management, which was predicted to occur at intermediate GLR values, among other

benefits.

3.2.4.3. Topside compact separator I-SEP by Caltec Ltd.

A different approach to severe slugging mitigation with phase separation was undertaken by

Caltec Ltd. They propose to use an in-line compact uni-axial cyclonic separator and after a

choke valve installed topside.

A paper by Jones et al. [56] from 2014 reports a series of experimental trials aimed at

investigation of severe slugging mitigation capabilities of I-SEP® (compact cyclonic gas/liquid

separator by Caltec Ltd.). The experimental campaign conducted at the Cranfield University,

utilized its three-phase flow test facility using water and air.

The comparative analysis between the topside choking and combined choke/I-SEP® severe

slugging mitigation techniques was conducted. The choking/I-SEP® combination proved to be

Figure 3.13. Caisson-ESP concept [52].

a) Schematic depiction of field development with use of C-ESP

b) Simplified drawing of a C-ESP

a) b)

42

a more efficient flow stabilizer compared to standard topside choking, reducing the pressure

oscillations by up to 75% at the same pressure drop over the test facility [56]. Moreover, in

passive regime I-SEP® reduces the pressure fluctuations at the riser base by 60%, whereas the

flow is stabilized completely at higher relative choke opening and lower riser base pressure, see

Figure 3.14.

The authors point out that there is no rigor quantitative explanation to the effect of the I-SEP®

on severe slugging [56]. It well may be that the separator applies additional retardation force

during the blowout phase further stabilizing the flow in a manner comparable to choking,

however utilizing different mechanism. I-SEP® is more sensitive to a perturbation in flow rate,

resulting in greater stability benefit, thus there is active control involved in the use of I-SEP®.

3.2.5. Foaming

Hassanein and Fairhurst [57], discussing challenges related to mechanical and hydraulic design

of risers in deep water, pointed out that with increased water depth occurrence of bigger slugs

was expected owing to increase in flow line diameter. They proposed the use of foams as a

means to control severe slugging [57]. However, no further information regarding this technique

followed, neither experimental nor theoretical.

Figure 3.14. Standard choking compared to choking/I-SEP® bifurcation diagrams [56].

43

In 2014, Sarica et al. [58] published the results of the first experimental investigation of the

severe slugging mitigation using foaming. The foam was generated inside of the pipeline near

to the riser base by injecting surfactant into the flow.

The experiment was conducted on the experimental facility at the University of Tulsa aimed at

modeling severe slugging, which consisted of 20 m long pipeline section with negative

inclination of 3° followed by a 15 m riser. Inner diameter of the pipeline was equal to 3 in. Air

and water were used as testing fluids. Heavy Duty Foaming Agent SI-403 was used as foaming

agent [58]. It is a blend of surfactants that is used for unloading water from depleted gas wells

and is capable of forming stable foams under various conditions and phase composition of the

transported fluid. The injection point was located at the end of the pipeline, near to the riser

base.

A number of tests were run without surfactant injection to contour steady and transitional

slugging (SS2 and SS2) envelopes, followed by test with the foaming agent added to the flow

at riser base. Figure 3.15 maps all the trials.

Figure 3.15. Severe slugging map comparison [58].

Diamonds, squares and triangles represent results of the base case, whereas circles the

trials with foaming agent injected. The thick lines contour the efficient foaming region.

44

Addition of surfactants shifted the severe slugging curve to the left, assuring stable or near to

stable flow at lower superficial gas velocities, approximately halving the severe slugging region

[58].

Figure 3.16 gives a closer look at one of trials completely eliminating severe slugging. At some

point of the experiment, the pressure fluctuations in the riser drops almost to zero as it is shown

by the blue dots, indicating establishment of stable flow. In addition to that, the total pressure

loss over the riser length was significantly reduced compared to the average value between the

maximum and minimum of the pressure fluctuations, which is typical for other anti-slugging

techniques [58]. Therefore, implementation of the surfactants, in addition to its effect of

slugging reduction, brings along reduction in backpressure on the pipeline.

Independent of the amount of the foaming agent injected into the flow, surfactants cannot

eliminate slugging for some combinations of liquid-gas flowrate, as it may be seen on Figure

3.15 (blue dots with red rim). Moreover, surfactants may cause detrimental effects on the

pressure fluctuations in some cases, slightly increasing its value. That is related to improved

sweeping of the liquid during the blowdown phase caused by the presence of the surfactant [58].

Figure 3.16. Severe slugging elimination case [58].

The purple point on the right part of the figure indicates the case position on the slugging

map. The blue dots depict the value of the pressure fluctuation over one cycle. Notice how

the mean value of the pressure drop is lower with surfactant added compared to no

injection case (Step 1).

45

An overview of all the trials with respect to their effect on slugging is mapped by Figure 3.17.

The measure of reduction in pressure fluctuations is expressed by dimensionless elimination

performance index (EPI), which is expressed as [58]:

𝐸𝑃𝐼 = [1 −(∆𝑃𝑚𝑎𝑥 − ∆𝑃𝑚𝑖𝑛)𝑎𝑓𝑡𝑒𝑟 𝑖𝑛𝑗.(∆𝑃𝑚𝑎𝑥 − ∆𝑃𝑚𝑖𝑛)𝑏𝑒𝑓𝑜𝑟𝑒 𝑖𝑛𝑗.

]. (3.11)

It worth mentioning here, that in Sarica et al. [58] all the experiments were conducted with

constant pressure at the riser outlet. For that reason pressures are expressed as pressure drop

over the riser (∆𝑝). In general case, one would see the same pattern of pressure fluctuations and

the formula above also applies to pressure at its absolute value.

The graph above shows that EPI has strong correlation with both gas and liquid superficial

velocities. It may be inferred that decrease in superficial liquid velocity causes increase of slug

elimination efficiency, whereas reduction of superficial gas velocity has an opposite effect.

Moreover, Figure 3.15 shows that superficial gas velocity is the crucial factor for the efficiency

of the technique since the contraction of severe slugging region occurs along the gas velocity

axis. Injection of additional amounts of gas would move the operational point to the right, see

Figure 3.15, when the resulting amount of injected gas would be sufficiently lower compared to

conventional gas lift.

Figure 3.17. EPI of gas and liquid superficial velocities [58].

46

Based on the arguments above we can come to a conclusion that surfactant injection efficiency

could be enhanced by injection of a relatively small amount of gas. To address to that point, a

series of following trials on feasibility of combination of foaming and gas injection was run by

Sarica et al. [36]. The experiment was conducted on the same facility, using air and water with

rates close to the previous investigation and the same surfactant. In addition, possibility to inject

gas at the riser base was allowed. In total, 30 tests (combinations of gas and liquid superficial

velocities before injection) were run, each consisting of several steps with different gas and

surfactant injection rates.

The results of the experiment are plotted on Figure 3.18. Surfactant injection rate was always

above critical concentration.

Figure 3.18. Severe slugging elimination with foam and gas injection [36].

Each test consisted of 7 steps:

Step 1 – no injection (used to determine the base line);

Step 2 (Only Gas Injection) – maximum gas injection applied;

Step 3 – gas injection with surfactant;

Step 4-6 – gradual reduction of gas injection.

Step 7 (Only Surfactant) – gas injection stopped.

47

As it shown in Table 3.1 and figures above, addition of the surfactant to the flow characterized

by slugging with pressure fluctuations of about a half of the base case reduced pressure

fluctuations almost to zero (𝐸𝑃𝐼 = 1). Further reduction of the gas injection rate showed results

similar to those in Sarica et al. [58].

Along with the flow conditioning effect, injection of the surfactant decreased backpressure on

the pipeline. However, the measure for backpressure reduction (BPR) provided by Sarica et al.

[36] takes into account the maximum value of the pressure fluctuations instead of its mean. This

choice seems questionable.

To the downsides of the technique undesired foaming inside of the separator may be attributed.

To mitigate foaming an additional chemical must be added. There is also some evidence that

addition of surfactants and consequent addition of antifoaming agents caused flow assurance

problems related to increased proneness to wax deposition [59].

Table 3.1. Severe slugging elimination with

foam and gas injection [36].

Note: color coding of the EPI columns:

Blue – severe slugging eliminated (EPI>0.5);

Yellow – partly severe slugging elimination (0<EPI<0.5);

Red – detrimental effect of the pressure fluctuations (EPI<0).

48

Criticism.

Some of the tests provided by Sarica [36] give opposite results despite their similarity. Let us

take a closer look at Tests 3 and 4 in Table 3.1. Both have comparable initial gas and liquid

superficial velocities and almost identical gas and surfactant injection rates. Both show

comparable results in Steps 1-6. However, in the case of slugging mitigation with foaming only,

the two tests show completely different results. Test 3 has almost no pressure fluctuations,

meaning that no severe slugging is present, whereas the situation in Test 4 got worse compared

to the base case. Figure 3.15 can be used to determine the predicted flow regime. Both of the

cases lay in transitional severe slugging region and slugging should have been completely

eliminated by foaming. This could mean either a misprint, a mistake in the experiment, or

incomplete understanding of the process. The authors provide no comments on the result under

discussion.

The choice of the criteria to measure the effect on the backpressure reduction also seems to be

questionable. It takes into account the change between the maximum exerted pressures during

fluctuations instead of its mean values. The mean value of the pressure seems to be a better way

to describe the backpressure created by slugging.

3.2.6. Multiphase riser base lift

The multiphase riser base lift concept proposed by Johal et al. [60] was aimed to overcome the

main drawbacks of the standard gas lift method, especially for deep water installations. As it

was described above, to those one can attribute the necessity for high pressure gas supply and

Joule-Thompson cooling effect, resulting in increased possibility of hydrate and wax deposition.

The last is especially prominent for deep waters due to higher hydrostatic pressure difference,

therefore higher expansion of the rising gas inside of the riser.

The authors proposed to use a cross over path between the existing flowlines to reroute some of

the produced multiphase flow near the riser base. Presumably, it would allow to establish a

desired flow regime without supplying external energy to the system. It was highlighted that the

system is especially effective for start-up of operations, which may be considered to be the worst

situation, and could be more attractive in economic terms compared to the gas lift system [60].

49

The system requires a high capacity multiphase line available at the site and therefore is limited

only for specific cases [19].

50

4. Mitigation methods modelling

4.1. Case study construction

For the purpose of comparison between different mitigation methods and further evaluation of

the proposed method, a case study needs to be chosen. However, the author was unable to find

an example from the industry, which would fit all of the criteria. For that reason, an artificial

case study was modelled. This part describes the model and the main principles that were used

to construct it.

4.1.1. The pipeline-riser system

The overall pipeline geometry is presented in Figure 4.1. and summarized in Table 4.1. It is

comprised of 5.9 km pipeline section followed by 500 meters high vertical riser. At the point of

1 km downstream of the wellhead a 250 meters long pipeline section with negative inclination

of 2° is situated. It is followed by an identical section with positive inclination, bringing the

pipeline back to its zero level. The two sections are thought to simulate some of the hilly-terrain

effects. The riser outlet is followed by 50 meters horizontal section according to the manual

recommendations [61]. Moreover, some evidence in literature was found that absence of the

horizontal section prior to separator may result in unphysical results, primarily due to excess of

fallback liquid [62].

51

Table 4.1. Pipeline geometry

Pipeline section Length,

[m]

Inclination,

[grad]

Inner diameter,

[inch]

Roughness,

[mm]

1 1000 0

6 0,03

2 250 -2

3 250 2

4 4000 0

5 400 -3

Riser 500 90

Horizontal pipe 50 0

4.1.2. The fluid model

As a descriptive model for the fluid properties the Blackoil module within OLGA was chosen.

The Blackoil model is appropriate for computations when little information is available about

the production fluid. Oil, gas and water specific gravities, and gas-oil ration (GOR) are the only

parameters required for the model [61], making it a perfect choice for the limitations imposed

by the scope of the thesis.

modelMassConvRef : modelPipelineRefgfedcb

x [m]

5 5005 0004 5004 0003 5003 0002 5002 0001 5001 0005000

y [m

]

460

440

420

400

380

360

340

320

300

280

260

240

220

200

180

160

140

120

100

80

60

40

20

0

-20

Figure 4.1. The study case pipeline geometry generated in OLGA.

NB! The axises have different scale.

52

A PVT table is not required when Blackoil model is used. Instead, at each instance of time and

at every position along the pipe the module uses predetermined correlations to calculate

equilibrium between oil and gas, assuming that gas is dissolvable in oil, but no oil can exist in

the gas as gas phase. Based on that, physical properties of fluids are calculated.

Glasø correlations were chosen to model gas-oil dissolvability. It is advised to use for oils of

North Sea region of API greater than 15. The bubble-point pressure may be acquired implicitly

from the equations [61]:

{log (

𝑃𝑏𝐶𝑃𝑏

) = 1.7669 + 1.7447 log(𝐴) − 0.30218[log(𝐴)]2,

𝐴 = (𝐺𝑂𝑅/𝛾𝐺)0.816(𝑇0.172/𝐴𝑃𝐼0.989).

(4.1)

For given pressure and temperature combination the equilibrium 𝑅𝑆𝐺𝑂 may be calculated as

follows [61]:

𝑅𝑆𝐺𝑂 = 𝐶𝑅𝑆𝐺𝑂𝛾𝐺 [𝐴𝑃𝐼0.989/𝑇0.172 ∙ 10

0.5 log(𝑃)−1.3292.6256 ]

−0.816

. (4.2)

It is important to notice that the model units are psia and °F.

The oil density is primarily a function of pressure, temperature and 𝑅𝑆𝐺𝑂. At pressures higher

than bubble-point pressure, 𝑅𝑆𝐺𝑂 = 𝐺𝑂𝑅, and the density follows regular isothermal

compressibility rule:

𝜌𝑂 = 𝜌𝑂𝑏 exp(𝑐𝑂(𝑃 − 𝑃𝑏)), (4.3)

where 𝑐𝑂 is the compressibility coefficient and may be calculated as [61]:

𝑐𝑂 = 10−5(−1433 + 5𝑅𝑆𝐺𝑂 + 17.2𝑇 − 1180𝛾𝐺 + 12.6𝐴𝑃𝐼)/𝑃. (4.4)

For pressures below the bubble-point, the oil density is computed in the following way [61]:

𝜌𝑂 =62.4𝛾0 + 0.0136𝛾𝐺𝑅𝑆𝐺𝑂

0.9759 + 1.2 ∙ 10−4[𝑅𝑆𝐺𝑂(𝛾𝐺/𝛾0)0.5 + 1.25𝑇]1.2. (4.5)

The correlation above may be also used to compute oil density at the bubble-point by

substitution 𝑅𝑆𝐺𝑂 with 𝐺𝑂𝑅.

The model units for pressure, temperature and FVF are psia, °F, and bbl/STB, respectively.

53

The gas density is calculated with assumption of the real gas model, where compressibility

factor 𝑧 is precalculated and stored within the software. Phase viscosities and interphase surface

tensions are calculated in the same fashion by using the model correlations.

The oil is assumed to have API of 30 and the produced gas specific gravity of 0.64.

Due to variety of flowrates modeled, the flow is assumed to be adiabatic. This was done to get

results independent of heat exchange with the environment, which in turn is highly dependent

on the flow regime, flow composition, and flowrate. Thus, by assuming adiabatic flow, heat

exchange was excluded from the model in hope to get more consistent results. To compensate

for that fact, the reservoir temperature was intentionally underestimated. A similar approach

was used in [63] where isothermal approximation was used. The authors advocate that severe

slugging effect is highly dominated by gravity forces and temperature variations typical for

subsea pipeline systems do not affect the results significantly.

Using the Blackoil model described above, the relation between bubble-point pressure and GOR

may be calculated and used to choose reasonable estimations of reservoir conditions, see Figure

4.2.

Figure 4.2. Bubble-point pressure dependence on initial GOR at 50°C (assumed reservoir

temperature).

54

4.1.3. The well and wellbore

Production volumes are assumed to be conducted through the 4½ in. production tubing, with

inner diameter of 3.958 in. [64]. Wall absolute roughness is set to 0.06 mm. The wellbore of

total length of 3 kilometers deviates from vertical line at 2 kilometers depth. The last kilometer

is drilled with 45° inclination.

The reservoir is assumed to be a saturated oil reservoir; therefore, Vogel’s IPR equation is used

to model the oil-well performance. Vogel’s equation reads as follows [61]:

𝑄0 = 𝑄0,𝑚𝑎𝑥 [1 − 0.2 (𝑃𝑤𝑓

𝑃𝑟𝑒𝑠) − 0.8 (

𝑃𝑤𝑓

𝑃𝑟𝑒𝑠)2

], (4.6)

where 𝑄0,𝑚𝑎𝑥 – maximum oil flow at flowing bottom hole pressure equal to zero.

Given the combination of true vertical depth from the mean sea level and desired GOR of 150,

a combination of reasonable reservoir pressure, flowing bottom hole pressure, and bubble-point

pressure were chosen. Thus, the reservoir exists at 315 bar of absolute pressure at 50 °C in

saturated condition. At flowrate of about 1500 Sm3/d the flowing bottom hole pressure

fluctuates around 270 bar. Maximum oil flowrate was assumed to be 6800 Sm3/d.

On importance of wellbore inclusion.

The system volume is one of the most important factors for severe slugging process due to gas

compressibility and its direct effect of the force balance described above. Meng and Zhang [65]

report erroneous flow assurance evaluation due to absence of a wellbore in the model studied.

Other researchers report severe slugging induced after connection of an additional well to a

flowline. Thus, despite increased flowrates, additional volume created conditions favorable for

terrain slugging. The model studied in this project has total volumes of pipeline and wellbore of

the same magnitude. Therefore, inclusion of a wellbore was considered to be necessary.

4.1.4. The topside facilities

As was briefly mentioned before, a riser model should include a horizontal section at the outlet.

Thus, the vertical riser part translates into 50 meters horizontal part and terminates into a

constant pressure node, which simulates operations of a separator. The pressure node is assumed

to operate at constant pressure and temperature. Exclusion of a separator simplifies the model.

Moreover, evidence was found that inclusion of a separator in OLGA models does not affect

55

severe slugging significantly. Thus, Ioannou et al. [66] reports that addition of a separator into

the model just slightly changed severe slugging behavior, affecting slugging frequency, but

leaving pressure fluctuations value unchanged, see Figure 4.3.

The constant pressure node serves as a boundary condition, allowing inflow of all phase and

outflow of only gas phase. It is made possible by setting GASFRACEQ node key (gas fraction

equilibrium) to 1, whereas WATERFRACEQ and OILFRACEQ to 0.

The separator pressure was set to 50 bar at normal operating conditions.

4.1.5. Solution convergence analysis

The base case assumed discretization method over the pipeline length with the rule of at least 2

sections per pipe and pipeline sections less than 100 meters. To investigate on the solution

convergence with respect to pipeline section length, sensitivity analysis was conducted with

discretization limits of 25, 50, 100, 200, and 400 meters. For all of the cases the well model was

substituted with constant mass flowrate source with gas flowrate of 0.84 kg/s and liquid flowrate

of 10.24 kg/s. At given mass flowrates transitional to severe slugging regime with stable slug

occurrence frequency was observed during preliminary studies. Readings of the total liquid mass

Figure 4.3. Pressure readings after the production choke for cases with and without

separator [66].

Simulations were done in OLGA7 with HD module. Delay constant was kept at 150.

Introduction of a separator affects slug frequency; pressure fluctuation remains the same.

56

flowrate at the riser outlet were used to characterize if riser-induced slugging was properly

simulated. Figure 4.4. (a) plots third slugs occurred in the system for each of the cases.

It may be inferred from the figure that resolutions of 200 and 400 meters provide numerical

solutions inconsistent with physical nature of the phenomenon. In addition to that, slug

Figure 4.4. Riser outlet liquid mass flowrate for different discretization resolutions.

Note: color coding of discretization rules (in addition to minimum number of

sections per pipe equal 2):

Black – MAX pipeline section length 25m;

Red – MAX pipeline section length 50m;

Blue – MAX pipeline section length 100m;

Green – MAX pipeline section length 200m;

Brown – MAX pipeline section length 400m.

a)

b)

57

frequency for the latter cases is consistently greater than for those with finer resolution. Closer

look at the fine resolution cases (Figure 4.4.(b)) shows that 100 and 50 discretizations have

flaws in solutions similar to those of rough discretization. The effect disappears if pipeline

section equal or less than 25 meters.

The drawback of having a fine meshing resolution is increased computational time. As a rule of

thumb, reduction of each section length by 2 translates into 4 times longer simulation time [61].

To maintain computational time within acceptable margins and benefit from a finer

discretization, progressive discretization was used. Thus, pipeline section lengths vary between

400 meters for long horizontal sections and 25 meters for the dip, pipeline-riser junction, and

riser. The ratio between adjacent pipeline sections was kept at about 1.5.

Resulting pipeline discretization encounters 71 pipeline sections, compared to 261 for the

meshing rule limiting sections length to 25 meters.

To test how well the refined discretization stands to that of 25 meters limiting factor generated

by OLGA standard tool, comparative simulation was run over 3 hours with the same flowrates

as described before. The result is presented on Figure 4.5. As one can see, the results are

comparable to a high degree. At the same time, significant amount of computational time was

saved due to reduced number of sections.

Investigation on the influence of the wellbore pipeline section length was conducted where

discretization length was reduced from basic 200 meters to 50 meters. Simulation showed

almost identical results, increasing computational time. Therefore, the wellbore section length

left to be 200 meters.

58

4.1.6. Flow regime map identification

Despite the fact that severe slugging occurrence criteria are beneficial for understanding of the

phenomenon mechanics, it is troublesome to use the same tool for real pipeline riser systems.

Thus, the common metrics used for severe slugging identification are superficial gas and liquid

velocities. In the laboratory conditions they may be accurately measured and controlled. Water

and air at PVT conditions close to standard are used as test fluids. Despite the fact that superficial

velocities may be the best choice in severe slugging effect description in laboratory conditions,

they are less suited for the task for a number of reasons:

1. The system of concern experiences great variety of pressure conditions, primarily

affecting the gaseous phase compressibility, hence its superficial velocity.

2. Unlike the behavior of gaseous and liquid phases of typical test fluids, gas is highly

dissolvable in oil meaning possibility of mass transfer between the phases.

An in-depth description of mass transfer effect on severe slugging was covered in Section 2.3.

Figure 4.5. Riser outlet liquid mass flowrate for progressive and fixed discretization.

Note: color coding of discretization rules (in addition to minimum number of

sections per pipe equal 2):

Black – MAX pipeline section length 25m;

Red – progressive discretization (section length=[25;400] meters);

59

For that reasons GOR and liquid production volume flowrate and standard conditions were

chosen to describe severe slugging occurrence in the system. To make the results comparable

with those from other sources, flow regime maps were also expressed in terms of volumetric

liquid and gas flowrates at standard conditions.

For the purpose of severe slugging operational conditions identification the model of producing

well was substituted with a constant mass flow source. The flowrate was tuned by changing

FEEDSTDFLOW key, set to change volumetric flowrate at standard conditions of the oil phase

(PHASE=OIL key) of the Blackoil modelled fluid.

A series of parametric studies was conducted with each of the simulation lasting for 8 hours,

with steady state pre-processor turned on (STEADYSTATE=ON key). First 2 hours of each

simulation were rejected due to oscillatory processes resulted from transition from steady state

pre-processor state to transient simulation. Remaining 6 hours were analyzed and flow regimes

identified and sorted into three categories:

1. Severe slugging of type 1;

2. Severe slugging of type 2;

3. Severe slugging of type 3;

4. Stable flow (a flow of relatively high frequency, no liquid production starvation phase).

The classification is consistent with the aim of the thesis, namely – severe slugging identification

and elimination. Transition from severe slugging state to slug flow was considered to be

sufficient. Therefore, no distinction between bubble and slug flow of relatively short slugs were

made and those two were combined on the flow map.

The flow regimes were primarily identified by readings of two graphs: mass liquid flowrate

(GLT trend variable) at the pressure (PT trend variable) at the riser base. Additionally, surge

liquid volume (LSLEXP trend variable specified at the riser outlet) and superficial gas velocity

at standard conditions at the riser base (USGST trend variable) were used for some flows.

GLT readings allowed verifying severe slugging presence and SS1 identification due to

characteristic feature of slow liquid production phase. SS2 and SS3 were differed by superficial

gas velocity readings at the riser base; for SS3 there is gas penetration at the riser base almost

at all times.

60

The resulting flow map outlining severe slugging occurrence region is presented on Figure 4.6.

The simulation summary is presented in form of tables in Appendix A.

Some peculiar results were observed in cases of low GOR values. Instead of severe slugging

with increasing period, starting from oil volumetric flowrate of 1500 Sm3/d at GOR equal to 50

steady flow regime with sudden release of fast liquid surges was present. This flow regimes

were not included on the flow regime map. The same phenomenon was observed for 40 bara

pressure in Section 5.4.

4.2. Topside choking

To model the effect of topside choking, the basic study case model was complimented with a

choke valve installed in the middle of the horizontal section before the separator. The following

settings were chosen based on the modelling purpose and the user manual recommendations:

MODEL = HYDROVALVE;

EQUILIBRIUMMODEL = HENRYFAUSKE;

THERMALPHASEEQ = NO;

SLIPMODEL = NOSLIP;

Figure 4.6. Severe slugging occurrence flow regime map for the separator pressure of 50

bara.

The green circles represent non-severe-slugging flow. The triangles, squares and crests

stand for severe slugging of type 1, 2 and 3, respectively.

61

RECOVERY = YES;

DIAMETER = 0.1524 [m];

CD = 0.84;

CR = 1.

The stepwise closing of the choke valve from fully opened position to 0.05 opening was

performed similar to that in [15]. Each step lasted for 3 hours, resulting in 21 hour simulation.

Time required to shift between the choke positions was set to 0.1 hour and relative opening was

changed linearly. The result is presented on Figure 4.7.

From here the relative choke opening is defined as a fraction between open cross sectional area

of partially actuated choke to its fully open inner cross sectional area, which for our case is equal

to inner cross sectional area of the pipeline.

62

Figure 4.7. Severe slugging mitigation with topside choking of relative valve opening

ranging from 1 to 0.05.

The graphs shows change in liquid mass flowrate into separator (b), pressure

fluctuations at the riser base (c), and total liquid production (d) with respect to

change in relative choke opening (a). Graphs are generated in OLGA.

a)

b)

c)

d)

63

The graphs show that the severe slugging effect is alleviated with increase of choking. Thus, at

relative opening of 0.1 a relatively mild slug flow observed, whereas at 0.05 opening the flow

is fully stabilized. However, the effect comes at cost of total production rate. The initial

production rate, fluctuating between approximately 1460 and 1700 Sm3/d and having arithmetic

average of 1627 Sm3/d, is reduced to 1470 Sm3/d in the case of fully stabilized flow due to

increased backpressure on the producing well.

A series of consecutive studies with shorter step of closing ratio was conducted to give a closer

look at the flow behavior at low relative choke openings, which summarized on Figure 4.8. As

before, the more aggressive the choking effect was, the better flowrate was stabilized at the cost

of lowered production and increased back pressures on the production wells.

The study intentionally does not provide conclusion on which closing ratio must be preferred.

The choice is case sensitive and highly depends on the capabilities of the process facilities

available. Thus, if the first stage separator has the maximum liquid throughput of 3400 Sm3/d

(approximately double the maximum well production in the initial case) and the system may

tolerate surges of up to 3 Sm3, the optimal choice of relative choke opening would be 0.12 as it

is shown on Figure 4.9.

64

Figure 4.8. Severe slugging mitigation with topside choking of relative valve opening

ranging from 0.24 to 0.06.

The graphs shows change in liquid mass flowrate into separator (b), pressure

fluctuations at the riser base (c), and total liquid production (d) with respect to

change in relative choke opening (a). Graphs are generated in OLGA.

a)

b)

c)

d)

65

The effect of topside choking with respect to pressure fluctuations at the riser base may be

neatly summarized under a common bifurcation diagram depicted on Figure 4.10. The

bifurcated part represent maximum and minimum values of pressure gauge reading with the

average value between plotted as a dash line. At relative opening between 0.06 and 0.07 the

flow is fully stabilized and the two curves merge into one.

Simulation results that allowed to plot the bifurcation diagram may be found in Appendix E.

Figure 4.9. Surge liquid volume of the system with maximum drainage rate of 3400 Sm3/d

with respect to relative valve opening.

The graph shows approximate solution to the problem of optimal choking choice

provided surge liquid capability of 3 m3. Relative valve opening of 0.12 satisfies

conditions of no carry-over. Graph is generated in OLGA.

66

4.3. Riser-base gas injection

To simulate severe slugging elimination with gas injection at the riser base, the base model was

altered by adding constant mass flowrate source at the riser base. The gas source was specified

to have constant injection temperature of 40 °C, using the Blackoil feed with GASFRACEQ node

key (gas fraction equilibrium) set to 1, whereas WATERFRACEQ and OILFRACEQ to 0,

allowing only gas injection with specified flowrate.

Based on the work of other authors, flowrates of the same order as the amount of gas produced

were expected to be necessary to stabilize the flow [6, 65]. The average gas flowrate at standard

conditions was estimated at approximately 243,000 Sm3/d. A gas injection rates of 50, 100, 150,

200, 250, 300, 350 and 400 thousand Sm3/d were investigated and compared to the base case.

The simulations results are summarized in Table 4.2. and Figure 4.11.

Figure 4.10. Topside choking bifurcation diagram

The graph shows absolute pressure fluctuations value dependence on the relative

choke opening. The upper limb represents pressure maximum, whereas the lower –

minimum. At relative opening of ~0.065 the flow is fully stabilized and the two lines

merge into one.

67

Table 4.2. Gas injection effect on severe slugging

Gas

injection

rate,

[103∙Sm3/d]

Maximum

liquid

flowrate

observed,

[kg/s]

Slugging

cycle

average

period,

[h]

Maximum

pressure at

riser base,

[bara]

Minimum

pressure at

riser base,

[bara]

Pressure

amplitude,

[bar]

Average oil

production,

[Sm3/d]

Average

pressure

at riser

base,

[bara]

Maximum

liquid

surge

volume,

[m3]

0 187 0,26 88,9 60,5 28,4 1624,8 74,6 5,1

50 128 0,19 82,6 59,7 22,9 1668,4 72,1 3,3

100 90 0,14 78,8 59,8 19 1698,9 70,3 2,2

150 77 0,12 77,2 58,5 18,7 1724,5 68,9 2

200 66 0,12 74,5 58,6 15,9 1748,7 67,3 1,2

250 58 0,1 73 58,9 14,1 1764,6 66,2 1

300 52 0,1 71,3 59,1 12,2 1776,9 65,5 0,5

350 37 0,08 68,5 59,8 8,7 1785,5 64,9 0

400 30 0,07 67,1 60,7 6,4 1795,5 64,3 0

It may be inferred from Figure 4.11.(a) that increase of the gas injection flowrate significantly

reduces pressure fluctuations at the riser base, at the same time reducing the average pressure.

As the result, the backpressure imposed on the producing well also decreases, boosting oil

production rate as it is shown on Figure 4.11.(b).

Figure 4.11. Pressure fluctuations (a) and average production rate (b) dependence on the

gas injection rate.

The left plot shows pressure fluctuations decrease with increased amount of injected

gas. As the result of gas lift the average production increases (right plot).

a) b)

68

4.4. Diameter reduction

Riser diameter reduction as a severe slugging alleviation technique, proposed by Yocum [18]

and calculated as a mitigation method in [65], was investigated on its effectiveness. To do so,

the case study model geometry was altered so that cross sectional area was equal to 0.9, 0.8 and

0.7 of the base case as it is shown in Table 4.3. These values were assumed to be reasonable

with respect to other operational considerations such as pigging.

Table 4.3. Riser diameter reduction effect on severe slugging

Area ratio

to the base

case

Riser inner

diameter,

[m]

Average cycle

period,

[h]

Period standard

deviation,

[h]

Maximum observed liquid

mass flowrate,

[kg/s]

Base

case 1 0.1524 0.227 0.024 187

Case 1 0.9 0.1446 0.259 0.044 175

Case 2 0.8 0.1363 0.303 0.038 184

Case 3 0.7 0.1275 0.372 0.071 177

Diameter reduction had almost no effect on the system behavior. Liquid surge volumes,

maximum liquid flowrates and total production remained at approximately the same level. One

notable change occurred to slugging period. In contrast to results in [65], riser diameter

reduction reduced severe slugging frequency. A possible explanation to that effect could be

increased superficial gas velocities resulting in improved liquid sweep out during blowdown

phase. However, no direct evidence favoring the explanation was found.

69

5. Severe slugging mitigation by backpressure reduction with a Surface jet pump

5.1. Proposal description

The majority of the severe slugging elimination techniques are based on the principle of increase

of backpressure on the pipeline-riser system, mainly by choking of the flow stream from topside

either in an automated or manual mode. The approach was thoroughly described in the previous

chapters. However, it may seem unreasonable to induce additional backpressure load on already

ageing production wells. Even though it may ultimately eliminate the problem of slugging, it

also decreases production rate and the lifetime of the field.

The method proposed here belongs to Caltec Ltd. and involves reduction of the backpressure

caused by processing facilities on a platform. It is achieved by outlet pressure reduction from

the first stage separator by means of an optional pump (-s) for the liquid phase and a surface jet

pump for the gaseous phase as it is shown on Figure 5.1.

The surface jet pump installed on the gas line after the first stage separator is deemed create a

pressure difference, hence lowering the separator outlet pressure while discharging gas at the

Figure 5.1. The layout of proposed method.

70

required downstream pressure. An optional pump for each of the liquid phase leaving the

separator may be also installed only if the reduced operating pressure is lower than downstream

pressures. As the result of the system alteration the first stage separator may be lowered.

5.2. Hypothesis

The setup is deemed to lower the overall backpressure load created by the topside facilities on

the subsea conduit for the cases where it may not be done in an alternative way, consequently

lowering the backpressure imposed on the producing well. Depressurized pipeline conditions

will allow the gas phase to expand, increasing its volumetric flowrate, thus increasing superficial

gas velocity, resulting in change of the flow regime inside of the pipeline from stratified flow to

stratified wavy flow and sweep-out of the accumulated liquids, dropping the pressure loss over

the transportation system even further. In addition to that, the well is supposed to respond on

decreased pressure with increased production rates, provided sufficient PI of the well, which

will further increase superficial liquid and gas velocities. Increased superficial velocities push

the operational point of the riser out of the severe slugging boundary, solving the problem of

severe slugging.

The primary effect from the method implementation should come from increased superficial gas

velocities and as a consequence possible transition from stratified flow to slug flow induced by

hydrodynamic mechanism. This in turn will effectively restrict severe slugging occurrence as

was previously discussed. The secondary effect comes from increased production rates due to

lowered backpressure on the system. There is another advantage in Brownfields, by lowering

the imposed backpressure some of the marginal backed-out wells which are tied-into some

common manifold would have chance to flow too. Thus bring additional production gains.

As an additional benefit, lowered operational pressure inside of the separator may have positive

effect, increasing the effectiveness of phase separation.

One of the major advantages for the proposed method is its relatively low CAPEX and OPEX

and possibility of making use of already existing on platform facilities.

71

5.3. Surface Jet pump technology

Ejectors, also known as jet pumps, eductors, injectors, or thermocompressors, are a type of

hydraulic pumps, which is operated based on the Venturi effect. The difference in the naming

primarily stems from the particular application, whether it is injection, pumping, compression,

creation of vacuum, etc. However, the underlying physical principle stays the same. Thus,

ejectors, operated with incompressible fluids are normally referred to as jet pumps or eductors,

whereas the terms ejector and injector used for compressible fluids [67].

Further reference to the technology as “surface jet pump” is deemed to eliminate ambiguity

when the term “jet pump” is referred to a downhole artificial lift technology.

The device has no moving parts and allows mixing of two fluids, one of high total pressure

(𝑃𝐻𝑃) and of low total pressure (𝑃𝐿𝑃), discharging a fluid at medium pressure (𝑃𝑀𝑃), so that

𝑃𝐻𝑃 > 𝑃𝑀𝑃 > 𝑃𝐿𝑃. Due to the fact that the intermediate pressure is always greater than the

pressure at inlet, the device can be used for raising pressure.

5.3.1. History and applications

The ejector was invented by Henry Giffard in 1858 and found its first application in steam

locomotives, where the steam was used as a motive fluid to pump the boiler feedwater into the

boiler [68]. A part of high pressure steam from the boiler was used as motive fluid. In the

combining cone or mixer chamber, steam would condense, interacting with cold water, see

Figure 5.2. Surplus of steam is flushed via overflow outlet. Slightly heated and pressurized water

after delivery cone had enough pressure to overcome control valve and enter the boiler.

At the beginning of the 20th century, the principle was also used for removing air from a steam

engine condenser. Later, ejectors were put into use to create vacuum and for cooling systems.

Thus, in the early 1930s the technology was widely spread as a way for air conditioning of large

buildings [69].

Steam jet pumping is used in desalination systems to eliminate mechanical vapor compression

cycle, which uses high grade electrical energy. Jet ejectors proved to be the simplest, cheapest,

and most reliable systems for vapor compression compared to other methods of compression

making use of low grade heat energy [70].

72

Nowadays, injectors are found in power, chemical and refinery plants as well as in aerospace,

refrigeration and vacuum applications, reliably fulfilling their main target – injection or sucking

out.

At the same time, eductors remain the only alternative for pumping of high abrasive mixtures,

due to high resistance to erosion. It allows eductors to be used for bulk handling of grains or

other granular or powdered materials, flushing and pumping of slurries.

Ejectors are also widely used in wet scrubbers, which are an effective way of gas cleaning. The

same principle is in place; the motive fluid, however, accomplishes two purposes at the same

time: it creates suction for the gas to be cleaned and as the way to entrain particles or gases.

Apart from that, ejectors are used for small bubbles creation for industrial needs [71].

5.3.2. Principle of operation

As it was mentioned above, injectors are based on the Bernoulli principle. They are used to

pump or compress the secondary fluid by transfer of momentum and energy from the primary

fluid.

Motive fluid, having high pressure and low velocity, is accelerated to high velocity jet through

a converging or converging-diverging nozzle, where the initial potential energy of motive fluid

Figure 5.2. Steam locomotive boiler injector [72].

One of the first implementations of the jet pump and a beautiful example of smart

engineering. Steam was used to drive feed water into the boiler.

73

is converted into the jet momentum at the nozzle exit according to Bernoulli equation. Thus, the

escaping from the nozzle jet has high velocity and low static pressure and accelerates the

secondary fluid in the direction of the stream. The two fluids are mixed in the mixing tube.

Finally, the diffuser converts kinetic energy of the mixed fluids back to static pressure [67].

Let us have a look at an example of energy conversion according to Bernoulli equation in the

nozzle. For the sake of simplicity, we consider a horizontal injector, operated with

incompressible motive and driven fluids. No energy losses, such as losses in the nozzle or due

to friction, are present in the system. Then, for the flow of HP fluid before and after the nozzle,

Bernoulli equation may be written as:

(𝑃

𝜌+1

2𝑉2)

1

= (𝑃

𝜌+1

2𝑉2)

4

, (5.1)

where indexes 1 and 4 correspond to cross sections 1 and 4 on Figure 5.3, respectively. Due to

smaller diameter at the nozzle exit the flow speed at point 4 should be greater. Consequently,

the pressure at the same point should drop to keep the equation in balance. Thus, on the right

side of the nozzle the sucking pressure is created.

Due to reduced pressure and the action of friction forces, the driven fluid is entrained in the

flow. Momentum transfer occurs and the two fluids are mixed in the mixing tube. In the diffuser

the kinetic energy of the mixed fluids is converted back to potential, i.e. pressure. Thus, in the

outlet of the ejector we have pressure greater than initial pressure of LP fluid.

Compressible fluid ejector based on the same principle as described above. However, the

calculations become more difficult and effects, such as choking phenomena, compressibility,

and energy balance equation should be taken into account.

5.3.3. Typical construction

A typical injector consists of inlets for motive (high pressure) and driven (low pressure) fluids,

motive fluid nozzle, mixing tube and diffuser, see Figure 5.3. The mixing tube may be

represented by a constant-area mixing shape, as it is depicted on Figure 5.3, or a constant-

pressure mixing shape, where the mixing tube has conical shape with the wider side pointing

towards the nozzle, see Figure 5.4 [69].

74

In case of the motive fluid being a gas, the nozzle may utilize de Laval design, so called

converging-diverging nozzle, allowing the gas achieve speeds greater that the speed of sound in

given conditions. Injection of the motive fluid with supersonic speed allows a greater conversion

of primary fluid energy to secondary pressure increase [67]. For incompressible fluids, simple

converging nozzles are used.

Figure 5.3. Velocity and pressure distribution along the stream axis of a jet pump [73].

Figure 5.4. Different principles of mixing tubes (a,b) and design of an SJP (c) [69].

(c) Surface Jet Pump

75

5.3.4. Applications in oil and gas industry

Surface jet pumps have a number of significant advantages making them an important option

for compression or boosting purposes. To those one may attribute [74]:

1. No moving parts;

2. Little or no maintenance;

3. Reliable operation;

4. Easy in installation and controlling;

5. Low associated costs;

6. Low weight and size;

7. High safety level;

8. High tolerance to presence of multiphase flow, including solids;

9. Possibility to utilize otherwise wasted energy from HP sources;

10. Possibility to use liquid or gas as power fluid.

The main drawbacks of the technology are:

1. Relatively inefficient methods, with mechanical efficiency peaking at approx. 30%;

2. Significant volumes of power fluid required;

3. Performance may drop significantly with presence of multiphase flow;

4. Power fluid mixes with driven fluids (could be an advantage for mixing duties)

It is worth noticing that jet pump’s lower mechanical efficiency should not be overemphasized,

since the technology may harness already available sources of HP fluids or utilize existing

facilities.

The area where jet pumps are applied the most within the petroleum industry is their use for

artificial lift of oil. The jet pump is insensitive to its orientation in space and therefore can be

successfully applied in deep, directionally drilled, and crooked wells [75]. Absence of moving

parts, and relatively small dimensions of the pump make maintenance infrequent and

inexpensive. In addition, casing type installation allows a jet pump retrieval without pulling out

of the tubing. Instead, the pump may be pulled by reverse flow or by introducing a wireline. The

jet pump has high tolerance to corrosive and abrasive produced fluids as well as low

requirements for pureness of motive fluid (in terms of solid particles). Moreover, water can be

76

potentially used as the motive fluid. For that reasons, jet pumps are one of the best solutions for

highly deviated deep wells producing heavy oils [75].

The jet pump found some of its implications in petroleum processing. Thus, it is the only

possible solution to create vacuum for dearation purposes, start-up of operations, or sea water

handling [76]. Implementation of a jet pump may also effectively prevent gas flaring. After

many stages of separation, resulting gas pressure is often too low to be efficiently gathered. In

that cases, the gas is usually burned using gas flares. As an alternative, low pressure gas may be

compressed by means of a jet pump and rerouted directly to the export line or to processing

facilities. Either gas or liquid may be used as a motive fluid for flare gases due to small rates

[73].

Especial interest present the perspective of oil and gas production boosting with use of the jet

pump either installed on a platform or subsea. Various sources of motive fluid may be utilized

for that purpose.

Thus, Caltec [77] reports two implementation of Surface jet Pump on Hewett field between

1987 and 1991. As a result of utilization of available gas from HP producing wells, production

from LP gas wells was increased from 100 MMScf/d to 125 MMScf/d for Lower Bunter wells,

and from 36 to 51 MMScf/d for Upper Bunter wells.

Caltec [73, 78] refer to Gemini gas field, where a surface jet pump installed on the platform

used a compressor recycle line as a source of motive gas. It resulted in reduced backpressure

from the platform by 200 psi, which increased gas production from three subsea satellite wells

27 miles away. Increased flow rate swept away liquids accumulated in the pipeline, causing

further drop of 140 psi. As the result, production increased by 24%; 2.5 BScf of otherwise lost

reserves recovered because of the lowered backpressure.

Oil wells production boosting is somewhat more complicated and often involves

implementation of the inline separator (I-SEP by Caltec).

Caltec [78] reports that separated liquid from a high pressure oil well with additional boost from

an oil pump was used as a motive fluid in the WELLCOM system to increase production from

some oil wells. As the result, backpressure on the wells dropped by 11 bar, increasing oil

production 350 bbl/d and gas production by 1.5 MMScf/d.

77

Caltec [78] reports implementation of the WELLCOM system to bring some closed low pressure

oil wells back to production making use of the energy from a HP well. It resulted in increased

oil production by more than 35% (about 150 bbl/d).

5.4. Effect on a constant mass flowrate system

The main mechanism allowing severe slugging mitigation by pipeline internal pressure

reduction is based on the effect of gas phase compressibility. Thus, when the pressure is reduced

the gas phase expands, resulting in increased superficial gas velocity. If the velocity exceeds the

limit set by stratified flow criterion (see Section 2.4.1), the flow regime inside the pipeline shifts

from stratified flow to stratified wavy flow and further to the point where hydrodynamic slugs

are induced, restricting liquid accumulation at the riser base and severe slugging initiation. The

secondary effect comes from the fact that fluids that are typical for oil and gas pipelines allow

mass transfer between themselves. Therefore, pipeline pressure reduction shifts the equilibrium

point between the liquid and gaseous phases of petroleum to the favor of gas, further increasing

superficial gas velocity inside the conduit.

To investigate on the effect of reduced pipeline pressure on severe slugging, a series of

simulations was run similar to that of Section 4.1.6 to get the flow properties and plot flow

regime maps. The total of 122 simulations were documented, all of which may be found in

Appendices A to D. Resulting flow regime maps are presented on Figure 5.5 expressed in terms

of oil production rate/GOR and oil/gas production rate at standard conditions.

Figure 5.5 and Figure 4.6 from Section 4.1.6 show how severe slugging occurrence region

diminishes with reduced backpressure imposed on the system due to induced hydrodynamic

slugging inside the pipeline. The effect is more noticeable in the region of medium liquid

flowrates (>500 Sm3/d) and gas-oil ratios (>50 Sm3/Sm3).

One of the most important factors influencing the efficiency of this approach is the pipeline

declination angle, which positively correlates with tendency of stratified flow regime being

present in the conduit. Therefore, one must expect to see lessening effect of the proposed

mitigation technique with increasing declination angle value since stratified flow will be forced

before the riser bend.

78

Figure 5.5. Severe slugging occurrence flow regime map for the separator pressure of 40

(a), 30 (b) and 20 (c) bara.

The green circles represent non-severe-slugging flow. The triangles, squares and crests

stand for severe slugging of type 1, 2 and 3, respectively. The squares with crosses inside

represent the cases when it was hard to distinguish SS2 from SS3.

The plots show how with decreased separator pressure the SS region shrinks.

a)

b)

c)

79

5.5. Effect on a system with producing well

An additional benefit from the pipeline depressurization is increased production due to lower

backpressure on the producing wells. Increased volumes of produced fluid will increase

superficial velocities, further affecting severe slugging occurrence criteria in a positive way.

The effect is highly dependent on the reservoir characteristic, especially PI value.

To illustrate the mitigation technique, the study case was altered in the way described in Table

5.1.

Table 5.1. Reservoir characteristics

Reservoir pressure, [bara] 280

Maximum production rate, [Sm3/d] 6800

Separator inlet pressure, [bara] 30

Gas-oil ratio, [Sm3/Sm3] 130

As before, the saturated oil reservoir exists at 50 °C. At the specified conditions the flowing

bottom hole pressure is at 242 bara with average oil production rate of 1540 Sm3/d, resulting in

PI of approximately 18 STB/d/psi. With given parameters flow regime inside of the riser may

be characterized as severe slugging of type 3 with pressure at the riser base fluctuations greater

than 20 bar. It is further assumed that implementation of a jet pump at the first stage separator

gas outlet allowed its inlet pressure to be reduced by 10 bar.

The resulting flow regime transition is showed on Figure 5.6. Depressurization of the pipeline

system with implementation of a jet pump caused operational point, which is denoted as a red

star, to move from severe slugging region (Figure 5.6.(a)) to stable flow region (Figure 5.6.(b)).

Figure 5.7 compares the time histories of the key parameters for severe slugging, where the

black curves represent the case before and the grey curves after pipeline depressurization. Figure

5.7.(a) shows that the riser base pressure after jet pump implementation became close to constant

value, indicating a stable flow regime, whereas the initial case exhibits pressure fluctuation of

more than 20 bar. Liquid mass flowrate graph (Figure 5.7.(b)) with the mitigation technique

being implemented shows arrival of slugs with consistent frequency, compared to severe

80

slugging liquid surges peaks depicted on the same plot. Finally, Figure 5.7.(c) shows that the

backpressure imposed on the producing well in the case of depressurization is close to constant,

translating into a constant production rate, compared to that of untreated case, which displays

high production flowrate variations due to effect of severe slugging. It is important to notice that

production rate for the case of stable pipeline-riser flow is higher than the average production

rate in case of severe slugging, resulting in the operational point shift.

Figure 5.6. Flow regime transition from the untreated case (a) to the depressurized system

(b).

The dashed line represent all possible flow combinations with given gas-oil ratio of 130.

Red stars correspond to gas-oil flowrates combination for 30 and 20 bara separator inlet

pressure.

a) b)

81

Figure 5.7. Surface jet pump implementation effect on a system with a producing well

expressed as time history of pressure at the riser base (a), mass liquid flowrate into the

separator (b), and the well production rate (c).

Black curves stand for the case before the jet pump implementation and the grey curves for

after. Graphs are generated in OLGA.

a) b)

c)

82

5.6. Surface Jet pump calculation example

This chapter provides a simplified ejector calculation procedure for choice of HP fluid

quantities, as can be found in “Surface Jet Pumps (SJPs) for Enhanced Oil & Gas production”

by Beg and Sarshar [74].

A number of assumptions are to be made:

Pressure drop over the first stage separator is equal and constant to 5 bar;

Temperature inside of the separator is equal and constant at 25 °C;

Gas from the first stage separator must meet the pressure requirement of 25 bara to be

fed into the last stage compressor;

The compressor performance is not sensitive to gas flowrate;

Oil and gas inside the first stage separator exist at equilibrium specified by inlet pressure

and separator temperature and described by Blackoil model (see Section 4.1.2);

Motive and driven gas have the same properties.

The case of Section 5.5 is considered. The surface jet pump is installed on the gas line between

the first stage separator and the last stage compressor to lower the outlet pressure from 25 to 15

bara, reducing the inlet pressure to 20 bara from initial 30 bara. Therefore, pressure difference

of 10 bar must be created by the surface jet pump to achieve the same pressure as in the base

case.

The production rate with the mitigation technique being installed was calculated before and

equal to 1780 Sm3/d, see Section 5.5. To estimate the amount of gas routed to the LP inlet of

the surface jet pump, solution gas oil ratio at the first stage separator conditions have to be

calculated. Assuming that the fluids exist at equilibrium specified by inlet pressure of 20 bar

and temperature equal to 25 °C, RSGO may be calculated according to Equation 4.2:

𝑅𝑆|𝑃=20𝑏𝑎𝑟𝑎;𝑇=25°𝐶 = 4.4 [𝑆𝑚3/𝑆𝑚3]. (5.2)

Therefore, the amount of gas fed into the LP line of the jet pump is equal to:

𝑄𝐺0 = 𝑄𝑂0(𝐺𝑂𝑅 − 𝑅𝑆) = 1780 ∙ (130 − 4.4) ≈ 224 ∙ 103 [𝑆𝑚3

𝑑]. (5.3)

83

Engineers’ Handbook in Surface Jet Pump [74] provides with a graphical way of Surface Jet

Pump performance estimation, see Figure 5.8. Assuming that the source of HP gas at 100 bara

is available (e.g. export line, compressor recycle line), the amount of HP required may be

estimated. As it follows from the figure, discharge pressure to LP and HP to LP ratios are

necessary to do so:

𝑃𝑑𝑖𝑠𝑐ℎ𝑎𝑟𝑔𝑒

𝑃𝐿𝑃=25 𝑏𝑎𝑟𝑎

15 𝑏𝑎𝑟𝑎≈ 1,67,

(5.4)

𝑃𝐻𝑃𝑃𝐿𝑃

=100 𝑏𝑎𝑟𝑎

15 𝑏𝑎𝑟𝑎≈ 6,67. (5.5)

Applying the values to Figure 5.8, LP to HP gas mass flow ratio may be estimated to be around

0.7. Thus, taking into account that the motive and driven gases have the same properties,

required volumetric flowrate of HP gas is equal to:

Figure 5.8. Ejector performance curves [74].

The curves show how the key ratios relate to each other. The black dashed line represents

the example of this part.

84

𝑄𝐺0,𝐻𝑃 =1

0.7∙ 𝑄𝐺0 = 320 ∙ 103 [

𝑆𝑚3

𝑑]. (5.6)

Therefore, about 320,000 Sm3/d of motive gas at 100 bara is required to implement the method

and eliminate severe slug flow at a given study case. The most probable gas sources would be

partly diverted export line gas, HP gas for injection wells, or gas from the last compressor

recycle line.

85

Conclusions This thesis had two main tasks to accomplish:

1) Provide a thorough theoretical overview of severe slugging phenomenon and techniques

for its mitigation (Chapters 2-4);

2) Evaluate a novel mitigation method proposed by Caltec Ltd., UK, based on application of

the Surface Jet Pump technology (Chapter 5).

The first task is of importance due to the scarcity of the literature on the topic. The main body

of knowledge is scattered in various papers over the last four decades, making its understanding

somewhat troublesome. The author hopes that this thesis will become a reference work for those

interested in the topic. The present work explains in detail the phenomenon mechanism and

classification, points out at some possible confusions that may arise, and highlights the

importance of recent publications, e.g. contributing mass transfer effect to the understanding of

severe slugging cycle.

It may be even more difficult to find information regarding severe slugging mitigation

techniques. The thesis provides a state of the art, an overview of all alleviation methods that

were possible to find, including patents in the field. Chapter 4 provides some guidance and

examples to simulation of some of those in the multiphase simulation software OLGA.

Chapter 5 describes a new severe slug mitigation method based on pipeline system

depressurization. Even though its basic mechanism is not novel and well understood, its

application may have been overlooked due to restrictions imposed by process facilities and the

impossibility to reduce the pipeline outlet pressure. The thesis provides a new option, showing

how a surface jet pump implementation may eliminate severe slugging in a cost-efficient manner

and further increase production rates, in contrast to the majority of mitigation methods used up

to date, which to some degree restricts production flow.

The chapter demonstrates that the method induces non-detrimental hydrodynamic slugging

inside the pipe by increased superficial gas velocities and avoids severe slugging initiation

process at the riser base.

86

As the result of simulations conducted, positive results were achieved showing that the proposed

mitigation method is feasible for flow stabilization at some flowrates, see Section 5.4. However,

the thesis has a limited scope, which mostly considers pipeline-riser system’s behavior under

reduced pressure. To get a thorough picture, restrictions imposed by the process facilities and

the reservoir must be included. Thus, undesirable effects such as increased water cut and

excessive GLR due to higher flowrate and migration in the reservoir should be taken into

consideration.

The second part of the thesis achieved its main goal – to show that severe slugging may be

avoided by pipeline depressurization with a surface jet pump – by providing a study case where

the mitigation technique proposed by Caltec Ltd. eliminated severe slugging flow, making

transition to a stable slug flow.

87

Recommendations for further work, self-evaluation Due to time constrains for this project and delays related to getting a version of simulation

software, it was not possible to model the effect of active controlled choking on the example of

a topside choke controlled by a PID controller; as it was planned initially. It would be also

beneficial, but very time demanding, to include frequency spectrum analysis of flow oscillations

based on Fourier transformation.

Some inconsistency with other authors was found when identifying different types of severe

slugging. Based on general description of the types, SS3 regimes were observed for high gas

flowrates for the simulations of the separator pressure of 30 and 40 bara, see Figure 5.5.(a-b).

Another improvement would include gas-lift mandrel node within OLGA model for riser-base

injection method instead of constant mass flowrate as it was done in the thesis.

The dip at the beginning of the pipeline was not necessary. Even though it was interesting to

observe, no actual documentation of observations is made in the thesis. Also it may have

interfered with the results, though in a very subtle way.

Recommendations for further work:

1) In-depth investigation on the effect of mass transfer. Comparative analysis between two

models one of which would include Blackoil modeled oil-gas solution, the other would

consist of a PVT table with two fluids of identical properties, but not allowing for mass

transfer to occur (refer to [63, 79]). Development of such a model if not too complicated

would help in active control methods design and tuning prior to installation.

2) Those mathematical models describing mass transfer effect on severe slugging assume

oil and gas existing at equilibrium at a given pipeline cross section [63]. They may be

further improved by developing a severe slugging model which takes into account

discrepancy between boiling and condensation of the gaseous phase out of oil. However,

it may be that the resulting model will improve results just marginally, whereas

increasing complexity significantly.

3) It may be possible to implement the method proposed by Caltec Ltd. and evaluate its

effect on a gas condensate pipeline.

88

Reference list

1. Bratland, O., Pipe Flow 2: Multi-phase Flow Assurance. [Electronic book] [last accessed

on 14 June 2015]. Available from: http://www.drbratland.com/free-book-pipe-flow-2-

multi-phase-flow-assurance/.

2. Schmidt, Ž., Experimental study of two-phase slug flow in a pipeline-riser pipe system.

PhD dissertation. University of Tulsa, 1977.

3. Schmidt, Z., J.P. Brill, and H.D. Beggs, Experimental study of severe slugging in a two-

phase-flow pipeline-riser pipe system. Society of Petroleum Engineers Journal, October

1980: p. 407-414.

4. Schmidt, Z., J.P. Brill, and H.D. Beggs, Experimental Study of Two-Phase Normal Slug

Flow in a Pipeline-Riser Pipe System. Journal of Energy Resources Technology, 1981.

103(1): p. 67-75.

5. Malekzadeh, R., R.A.W.M. Henkes, and R.F. Mudde, Severe slugging in a long

pipeline–riser system: experiments and predictions. International journal of multiphase

flow, 2012. 46: p. 9-21.

6. Schmidt, Z., D.R. Doty, and K. Dutta-Roy, Severe slugging in offshore pipeline riser-

pipe systems. Society of Petroleum Engineers Journal, February 1985: p. 27-38.

7. Linga, H., Terrain slugging phenomena. Some experimental results obtained at the

SINTEF Two-phase Flow Laboratory. in Proceedings of the 3rd International

Conference on Multi-Phase Flow, The Hague, Netherlands. 1987.

8. Tin, V. and M.M. Sarshar, An investigation of severe slugging characteristics in flexible

risers. in Proc. 6th Int. Conf. on Multiphase Production, BHRG, France. 1993.

9. Nemoto, R.H. and J.L. Baliño, Modeling and simulation of severe slugging with mass

transfer effects. International Journal of Multiphase Flow, 2012. 40: p. 144-157.

10. Taitel, Y. and A.E. Dukler, A model for predicting flow regime transitions in horizontal

and near horizontal gas‐liquid flow. AIChE Journal, 1976. 22(1): p. 47-55.

11. Bøe, A., Severe slugging characteristics; part i: flow regime for severe slugging; part

ii: Point model simulation study. in Selected Topics in Two-Phase Flow 1981. NTH,

Trondheim, Norway.

12. Taitel, Y., Stability of severe slugging. International journal of multiphase flow, 1986.

12(2): p. 203-217.

13. Montgomery, J.A., Severe slugging and unstable flows in an S-shaped riser. PhD thesis.

Cranfield University, 2002.

89

14. Pots, B.F.M., I.G. Bromilow, and M.J.W.F. Konijn, Severe Slug Flow in Offshore

Flowline/Riser Systems. SPE Production Engineering, 1987: p. 319-324.

15. Havre, K. and M. Dalsmo, Active Feedback Control as a Solution to Severe Slugging.

SPE Production & Facilities, August 2002: p. 138-148.

16. Wyllie, M.W.J., Apparatus for Inserting into a Conduit. UK Patent Application GB

2280460A. Issued on 1 February 1995.

17. DNV, Recommended Practice, RP O501: Erosive Wear in Piping Systems. 2007: Det

Norske Veritas.

18. Yocum, B.T., Offshore riser slug flow avoidance: Mathematical models for design and

optimization, in Society of Petroleum Engineers of AIME. 1973: London, England.

19. Sarica, C. and J.Ø. Tengesdal, A New Technique to Eliminate Severe Slugging in

Pipeline/Riser Systems, in SPE Annual Technical Conference and Exhibition. 2000,

Society of Petroleum Engineers: Dallas, Texas, USA.

20. Schmidt, Z., J.P. Brill, and H.D. Beggs, Choking can eliminate severe pipeline slugging.

Oil & Gas Journal, 1979. 77(46): p. 230-238.

21. Jansen, F.E., O. Shoham, and Y. Taitel, The elimination of severe slugging—experiments

and modeling. International Journal of Multiphase Flow, 1996. 22(6): p. 1055-1072.

22. Jansen, F.E. and O. Shoham, A Study on the Optimization of Severe Slugging

Elimination. Journal of Petroleum Technology, 1991.

23. Jansen, F.E. and O. Shoham, Methods for Eliminating Pipeline-Riser Flow Instabilities.

1994, Society of Petroleum Engineers.

24. Jansen, F.E., Elimination of severe slugging in a pipeline-riser system. MSc thesis.

University of Tulsa, 1990.

25. Mokhatab, S., B.F. Towler, and S. Purewal, A Review of Current Technologies for

Severe Slugging Remediation. Petroleum Science and Technology, 2007. 25(10): p.

1235-1245.

26. Hill, T.J., Gas injection at riser base solves slugging flow problems. Oil & Gas Journal,

1990. 26: p. 88-92.

27. Hannisdal, A., R. Westra, M.R. Akdim, A. Bymaster, E. Grave, and D.T. Teng, Compact

Separation Technologies and Their Applicability for Subsea Field Development in Deep

Water, in Offshore Technology Conference. 2012, Society of Petroleum Engineers:

Houston, Texas, USA.

28. Havre, K., K.O. Stornes, and H. Stray, Taming slug flow in pipelines. ABB review, 2000.

4: p. 55-63.

90

29. Courbot, A., Prevention of severe slugging in the Dunbar 16-in. multiphase pipeline, in

Offshore Technology Conference. 1996: Richardson, Texas, USA.

30. Jansen, B., M. Dalsmo, L. Nøkleberg, K. Havre, V. Kristiansen, and P. Lemetayer,

Automatic control of unstable gas lifted wells, in SPE Annual Technical Conference and

Exhibition. 1999: Houston, Texas, USA.

31. Hollenberg, J.F., S. De Wolf, and W.J. Meiring, A method to suppress severe slugging

in flow line riser systems. in Proc. 7th Int. Conf. on Multiphase Technology Conference,

Cannes, France. 1995.

32. Shell Global Solutions. Active Slug Control. [Internet] [cited: 12 March 2015]; Available

from:

http://www.petronics.ntnu.no/petronics_workshop_2004/Presentations/Company/Henk

es_Shell%20-%20Active%20Slug%20Control.pdf.

33. Kovalev, K., A. Cruickshank, and J. Purvis, The Slug Suppression System in Operation,

in Offshore Europe 2003. 2003, Society of Petroleum Engineers: Aberdeen, UK.

34. Barbuto, F.A., Method of eliminating severe Slug in multi-phase flow subsea lines.

United States Patent 5,478,504. Issued on 26 December 1995.

35. Tengesdal, J.Ø., Investigation of self-lifting concept for severe slugging elimination in

deep-water pipeline/riser systems. PhD dissertation. Pennsylvania State University,

2002.

36. Tengesdal, J.Ø., C. Sarica, and L. Thompson, Severe Slugging Attenuation for

Deepwater Multiphase Pipeline and Riser Systems, in SPE Annual Technical

Conference and Exhibition. 2003, Society of Petroleum Engineers: San Antonio, Texas,

USA.

37. Wyllie, M.W.J. and A. Brackenridge, A Retrofit Solution to Reduce Slugging Effects in

Multiphase Subsea Pipelines–The Internal Riser Insert System (IRIS), in Subsea

International Conference. 1994: London, UK.

38. Malekzadeh, R. and R.F. Mudde, A Modelling Study of Severe Slugging in Wellbore, in

North Africa Technical Conference and Exhibition. 2012, Society of Petroleum

Engineers: Cairo, Egypt.

39. Brasjen, B.J., J.H. Hansen, W. Schiferli, and S.P.C. Belfroid, Experimental Investigation

of Terrain Slugging Formation, Evolution and Potential for Mitigation, in 16th

International Conference on Multiphase Production Technology. 2013, BHR Group:

Cannes, France.

40. Brasjen, B., J. Veltin, and J.H. Hansen, Mitigation of Terrain Induced Slugging Using

Mixer Devices, in SPE Annual Technical Conference and Exhibition. 2014, Society of

Petroleum Engineers: Amsterdam, The Netherlands.

91

41. Arnaudeau, M. and J. Corteville, Method and device for stabilizing the flow of a two-

phase fluid. European Patent Office EP0034079 B1. Issued on 6 July 1983.

42. Makogon, T.Y. and G.J. Brook, Device for controlling slugging. United States Patent

Application Publication US 2009/0301729 A1. Issued on 10 December 2010.

43. Makogon, T.Y., D. Estanga, and C. Sarica, A new passive technique for severe slugging

attenuation. in Proceedings of the 15th international conference on multiphase

production technology. 2011. Cannes, France: BHR Group.

44. Shen, J.J.S. and H. Yeung, Apparatus for mitigating slugging in flowline systems. United

States Patent US 7,857,059 B2. Issued on 28 December 2010.

45. Xing, L., H. Yeung, J. Shen, and Y. Cao, A new flow conditioner for mitigating severe

slugging in pipeline/riser system. International Journal of Multiphase Flow, 2013. 51(0):

p. 65-72.

46. Xing, L., H. Yeung, J. Shen, and Y. Cao, Numerical study on mitigating severe slugging

in pipeline/riser system with wavy pipe. International Journal of Multiphase Flow, 2013.

53(0): p. 1-10.

47. Xing, L., H. Yeung, J. Shen, and Y. Cao, Experimental study on severe slugging

mitigation by applying wavy pipes. Chemical Engineering Research and Design, 2013.

91(1): p. 18-28.

48. De Almeida, A.R. and M. De Albuquerque Lima Gonçalves, Device and method for

eliminating severe slugging in multiphase-stream flow lines. United States Patent

6,041,803. Issued on 28 March 2000.

49. Almeida, A.R. and M. de A Lima Goncalves, Venturi for severe slug elimination. in

Proceedings 9th International Conference on Multiphase Production. 1999. Banff,

Canada: BHR Group.

50. Benetti, M. and M. Villa, Field Tests on VASPS Separation and Pumping System, in

Offshore Technology Conference. 1997, Society of Pertoleum Engineers: Houston,

Texas. p. 119-127.

51. Ju, G.T., H.S. Littell, T.B. Cook, M.H. Dupre, K.M. Clausing, E.E. Shumilak, W.

Schoppa, and W.A. Blizzard, Perdido Development: Subsea and Flowline Systems, in

Offshore Technology Conference. 2010, Society of Petroleum Engineers: Houston,

Texas, USA.

52. Iyer, S., P. Lang, W. Schoppa, Y.D. Chin, and A. Leitko, Subsea Processing at Parque

das Conchas (BC-10): Taking Flow Assurance to the Next Level, in Offshore Technology

Conference. 2010, Society of Petroleum Engineers: Houston, Texas, USA.

92

53. Eriksen, S., H. McLernon, and C. Mohr, Pazflor SSPS Project; Testing and Qualification

of Novel Technology: A Key to Success, in Offshore Technology Conference. 2012,

Society of Petroleum Engineers: Houston, Texas, USA.

54. Bybee, K., VASPS Prototype in Marimba Field - Workover and Restart. Journal of

Petroleum Technology, May 2006.

55. Peixoto, G.D.A., G. Ribeiro, P.A. Barros, M.A. Meira, and T.M. Barbosa, VASPS

Prototype in Marimba Field - Workover and Re-Start, in SPE Latin American and

Caribbean Petroleum Engineering Conference. 2005, Society of Petroleum Engineers:

Rio de Janeiro, Brazil.

56. Jones, R., Y. Cao, N. Beg, and C. Wordsworth, The Severe Slugging Mitigation

Capability of a Compact Cyclonic Gas/Liquid Separator, in The 9th North American

Conference on Multiphase Technology. 2014, BHR Group: Banff, Canada. p. 477-487.

57. Hassanein, T. and P. Fairhurst, Challenges in the Mechanical and Hydraulic Aspects of

Riser Design for Deep Water Developments. in IBC Offshore Pipeline Technology

Conference. 1998. Oslo, Norway.

58. Sarica, C., P. Begen, E. Pereyra, and G. Kouba, Feasibility of Surfactants as Severe

Slugging Suppression Agents, in The 9th North American Conference on Multiphase

Technology. 2014, BHR Group: Banff, Canada. p. 365-376.

59. Beg, N., Personal communication. 2015.

60. Johal, K.S., C.E. Teh, and A.R. Cousins, An alternative economic method to riser-base

gas lift for deep water subsea oil/gas field developments. in Offshore Europe

Conference. 9-12 September 1997. Aberdeen, Scotland.

61. SPT Group. OLGA User Manual. Version 7.1.1, 2014.

62. Mokhatab, S., Severe slugging in offshore production systems. 2010, New York, USA:

Nova Science Publishers.

63. Nemoto, R.H. and J.L. Baliño, Modeling and simulation of severe slugging with mass

transfer effects. International Journal of Multiphase Flow, 2012. 40(0): p. 144-157.

64. OILProduction.net. API Tubing Table. [cited: 20 April 2015]; Available from:

http://www.oilproduction.net/files/005-apitubing.pdf.

65. Meng, W. and J.J. Zhang, Modeling and Mitigation of Severe Riser Slugging: A Case

Study, in SPE Annual Technical Conference and Exhibition. 30 September–3 October

2001, Society of Petroleum Engineers: New Orleans, Louisiana, USA.

66. Ioannou, K., M. Widener, S. Kashou, and D. Estanga, Terrain Slugging In a Subsea

Tieback: When Predictions Fail, in 8th North American Conference on Multiphase

Technology. 20-22 June 2012, BHR Group: Banff, Canada.

93

67. Liao, C., Gas ejector modeling for design and analysis. PhD dissertation. Texas A&M

University, 2008.

68. Kneass, S.L., Practice and Theory of the Injector. 1898: Wiley.

69. Boumaraf, L., A. Lallemand, and P. Haberschill, Ejectors and their usefulness in the

energy savings. 2010, New York: Nova Science Publishers.

70. El-Dessouky, H., H. Ettouney, I. Alatiqi, and G. Al-Nuwaibit, Evaluation of steam jet

ejectors. Chemical Engineering and Processing: Process Intensification, 2002. 41(6): p.

551-561.

71. Transvac. Produced Water Treatment. [Internet] [cited: 3 August 2014]; Available from:

http://www.transvac.co.uk/produced_water_treatment.php.

72. Wikipedia. Injector. [Internet] [cited: 30 July 2014]; Available from:

http://en.wikipedia.org/wiki/Injector.

73. Beg, N., The applications of surface jet pump technology to increase oil & gas

production. 2011.

74. Beg, N. and S. Sarshar, Surface Jet Pumps (SJPs) for Enhanced Oil & Gas Production:

Engineers' Handbook. 2014, Cranfield, UK: Caltec Limited.

75. Noronha, F.A.F., F.A. Franca, and F.J.S. Alhanati, Improved two-phase model for

hydraulic jet pumps. in SPE Production Operations Symposium. 9-11 March 1997.

Oklahoma City, USA.

76. Carstensen, C., MOM202 Process Plants for Oil and Gas. Course taugth at the

University of Stavanger, 2010.

77. Green, A.J., K. Ashton, and A.T. Reade, Gas Production Improvements Using Ejectors,

in Offshore European Conference. 7-10 September 1993, Society of Petroleum

Engineers: Aberdeen, Scotland.

78. Sarshar, S.M.M., The Recent Applications of Jet Pump Technology to Enhance

Production from Tight Oil and Gas Fields, in SPE Middle East Unconventional Gas

Conference and Exhibition. 23–25 January 2012, Society of Petroleum Engineers: Abu

Dhabi, UAE.

79. Baliño, J.L., K.P. Burr, and R.H. Nemoto, Modeling and simulation of severe slugging

in air–water pipeline–riser systems. International Journal of Multiphase Flow, 2010.

36(8): p. 643-660.

94

Appendix A.

Table A.1. Flow regime identification for separator pressure of 50 bara.

Gas-oil-ratio,

[Sm3/ Sm3]

Absolute pressure reading

at the riser base, [bara]

Pressure

fluctuation

value,

[bar]

Severe

slugging

cycle

period,

[h]

Maximum

observed

liquid

mass

flowrate,

[kg/s]

Flow

regime

Maximum Minimum

𝑄𝑜0 = 250 Sm3/d

50 91,2 60,4 30,8 * 130 SS1

100 90,6 57,4 33,2 1,25 168 SS1

150 91,3 57,3 34 1,05 193 SS1

200 90,2 57,1 33,1 0,97 188 SS2

250 85 56,9 28,1 0,68 153 SS2

300 77,7 56,2 21,5 0,56 110 SS2

350 79,7 56,4 23,3 0,55 122 SS2

𝑄𝑜0 = 500 𝑆𝑚3/𝑑

50 91,3 63,9 27,4 1,63 120 SS1

100 91,2 60,7 30,5 0,81 163 SS1

150 91,2 58 33,2 0,52 197 SS2

200 87,4 58,2 29,2 0,37 173 SS2

250 82 57,3 24,7 0,33 226 SS2

300 77,7 56,9 20,8 0,3 123 SS2

350 75 57,1 17,9 0,25 102 SS2

𝑄𝑜0 = 1000 𝑆𝑚3/𝑑

50 91 64 27 0,9 119 SS1

100 90 63 27 0,3 165 SS1

150 90,5 60,1 30,4 0,29 195 SS2

200 85 60 25 0,21 163 SS2

250 81,2 58,5 22,7 0,26* 152 SS2-SS3

300 78 59 19 0.37* 130 SS2-SS3

350 71 59,1 11,9 - - stable

𝑄𝑜0 = 1500 𝑆𝑚3/𝑑

50 89,5 75,3 14,2 0,27 53 **

100 89,4 63,1 26,3 0,18 165 SS1

150 88,8 60,6 28,2 0,25 191 SS2

95

200 83,1 59,6 23,5 0,29* 84,7 SS2

250 71,3 61,4 9,9 - 54 stable

300 - - - - - stable

350 - - - - - stable

𝑄𝑜0 = 2000 𝑆𝑚3/𝑑

50 91 64 27 0,9 119 **

100 86 65 21 0,15 140 SS3

150 87,6 62,3 25,3 0,22* 168 SS3

200 67,6 67,2 0,4 - - stable

250 - - - - - stable

300 66,2 64,8 1,4 - - stable

𝑄𝑜0 = 2500 𝑆𝑚3/𝑑

50 91 64 27 0,9 119 **

100 86 68 18 0,15 115 SS3

200 68,8 68,2 0,6 - - stable

300 68,1 67,3 0,8 - - stable

𝑄𝑜0 = 3000 𝑆𝑚3/𝑑

50 91 64 27 0,9 119 **

100 78,3 74,5 3,8 - - stable

200 70,6 69,5 1,1 - - stable

* - flow exhibit irregular behavior; the value of period is either approximate or omitted.

** - seemingly stable flow with sudden release of hast liquid surges (mentioned in Section 4.1.6).

96

Appendix B.

Table B.1. Flow regime identification for separator pressure of 40 bara.

Gas-oil-ratio,

[Sm3/ Sm3]

Absolute pressure reading

at the riser base, [bara]

Pressure

fluctuation

value,

[bar]

Severe

slugging

cycle

period,

[h]

Maximum

observed

liquid

mass

flowrate,

[kg/s]

Flow

regime

Maximum Minimum

𝑄𝑜0 = 250 Sm3/d

50 81,3 49,9 31,4 1,94 144 SS1

100 81,6 47,1 34,5 1,25 186 SS1

150 81,8 47,4 34,4 1,14 226 SS1

200 73,9 46,8 27,1 0,73 166 SS2

250 72 46,7 25,3 0,74 118 SS2

300 65,2 45,8 19,4 0,5 102 SS2

350 64,7 45 19,7 0,5 148 SS2

𝑄𝑜0 = 500 𝑆𝑚3/𝑑

50 81,4 53 28,4 1,15 146 SS1

100 81,2 49 32,2 0,58 193 SS1

150 81,6 47,9 33,7 0,47 223 SS2

200 71,6 47,1 24,5 0,37 169 SS2

250 69,7 46,9 22,8 0,26 130 SS2

300 64,3 46,1 18,2 0,25 132 SS2-SS3

350 62,3 46,2 16,1 0,24 115 SS2-SS3

𝑄𝑜0 = 1000 𝑆𝑚3/𝑑

50 80,9 57,6 23,3 0,5 114 SS1

100 80,1 50,4 29,7 0,29 196 SS2

150 80,2 48,6 31,6 0,33 204 SS2

200 73,7 47,8 25,9 0,28 185 SS2-SS3

250 69,5 48,2 21,3 0,39 161 SS3

300 58,4 49,3 9,1 0,37 55 SS3

350 - - - - - stable

𝑄𝑜0 = 1500 𝑆𝑚3/𝑑

50 79,6 60,6 19 0,31 96 SS1

100 80 51 29 0,19 197,4 SS2

150 75,5 49 26,5 0,24 174 SS2

97

200 - - - - - stable

250 - - - - - stable

300 - - - - - stable

𝑄𝑜0 = 2000 𝑆𝑚3/𝑑

50 78,7 65,1 13,6 0,21 87,2 **

100 79 52,1 26,9 0,25 174 SS2-SS3

150 - - - - - stable

200 - - - - - stable

250 - - - - - stable

300 - - - - - stable

𝑄𝑜0 = 2500 𝑆𝑚3/𝑑

50 78,9 64,6 14,3 0,17 80 **

100 76,4 56 20,4 * 143 SS3

150 - - - - - stable

200 - - - - - stable

300 - - - - - stable

𝑄𝑜0 = 3000 𝑆𝑚3/𝑑

50 79,6 66,3 13,3 0,14* 77 SS3

100 - - - - - Stable

150 - - - - - Stable

* - flow exhibit irregular behavior; the value of period is either approximate or omitted.

** - seemingly stable flow with sudden release of hast liquid surges (mentioned in Section 4.1.6).

98

Appendix C.

Table C.1. Flow regime identification for separator pressure of 30 bara.

Gas-oil-ratio,

[Sm3/ Sm3]

Absolute pressure reading

at the riser base, [bara]

Pressure

fluctuation

value,

[bar]

Severe

slugging

cycle

period,

[h]

Maximum

observed

liquid

mass

flowrate,

[kg/s]

Flow

regime

Maximum Minimum

𝑄𝑜0 = 250 Sm3/d

50 71,3 39,7 31,6 1,98 171 SS1

100 71,9 37,1 34,8 1,21 221 SS1

150 71,8 36,5 35,3 1,09 248 SS2

200 59,8 35,4 24,4 0,65 133 SS2

250 54,6 35,5 19,1 0,56 146 SS2

300 50,4 34,8 15,6 0,4 107 SS2

350 49,3 34,4 14,9 0,34 105 SS2

𝑄𝑜0 = 500 𝑆𝑚3/𝑑

50 71,6 42,5 29,1 0,89 151 SS1

100 71,7 38,2 33,5 0,61 228 SS2

150 66,3 37,5 28,8 0,52 213 SS2

200 58,2 36,6 21,6 0,38 130 SS2

250 54,6 36,6 18 * 149 SS2

300 52,7 35,6 17,1 * 120 SS3

350 51,4 35,1 16,3 105 SS3

𝑄𝑜0 = 1000 𝑆𝑚3/𝑑

50 70,9 49 21,9 0,45 155 SS1

100 70,9 39,9 31 0,37 228 SS2

150 65,6 38,6 27 0,36 218 SS2

200 58,9 38,2 20,7 - 138 SS2-SS3

250 48,1 38,7 9,4 - 45 stable

300 - - - - - stable

𝑄𝑜0 = 1500 𝑆𝑚3/𝑑

50 69,9 49 20,9 0,27 138 SS1

100 69,5 39,8 29,7 0,25* 217 SS2

150 60,4 39,9 20,5 0,36* 141 SS3

200 - - - - - stable

99

300 - - - - - stable

𝑄𝑜0 = 2000 𝑆𝑚3/𝑑

50 67,7 49,9 17,8 0,18 136 SS1

100 68,4 41,6 26,8 0,37* 207 SS2-SS3

150 - - - - - stable

200 - - - - - stable

300 - - - - - stable

𝑄𝑜0 = 2500 𝑆𝑚3/𝑑

50 67,5 51,6 15,9 0,15 118 SS3

100 - - - - - stable

150 - - - - - stable

200 - - - - - stable

300 - - - - - stable

𝑄𝑜0 = 3000 𝑆𝑚3/𝑑

50 66 52,7 13,3 0,13 103 SS3

100 - - - - - stable

150 - - - - - stable

200 - - - - - stable

300 - - - - - stable

* - flow exhibit irregular behavior; the value of period is either approximate or omitted.

100

Appendix D.

Table D.1. Flow regime identification for separator pressure of 20 bara.

Gas-oil-ratio,

[Sm3/ Sm3]

Absolute pressure reading

at the riser base, [bara]

Pressure

fluctuation

value,

[bar]

Severe

slugging

cycle

period,

[h]

Maximum

observed

liquid

mass

flowrate,

[kg/s]

Flow

regime

Maximum Minimum

𝑄𝑜0 = 250 Sm3/d

50 61,9 27,2 34,7 1,96 210 SS1

100 62,1 28,5 33,6 1,35 245 SS1

150 50,2 25,7 24,5 0,87 175 SS2

200 47 24,6 22,4 0,67 140 SS2

250 38,3 24,4 13,9 0,44 105 SS2

300 37,7 23,4 14,3 0,43 120 SS2

350 36,2 23,7 12,5 * 65 SS2

𝑄𝑜0 = 500 𝑆𝑚3/𝑑

50 61,7 31,2 30,5 0,71 214 SS1

100 59,5 28 31,5 0,71 228 SS2

150 49,2 26,1 23,1 0,52 153 SS2

200 43,8 25 18,8 0,34 175 SS2

250 42,8 25 17,8 0,37 107 SS2

300 39,5 24,4 15,1 0,44 146 SS2

350 36,3 24,3 12 0,5 84 SS2

𝑄𝑜0 = 1000 𝑆𝑚3/𝑑

50 61,3 33,7 27,6 0,38 203 SS1

100 61,5 30 31,5 0,5 234 SS2

150 47,7 26,9 20,8 * 148 SS2

200 - - - - - Stable

250 - - - - - Stable

𝑄𝑜0 = 1500 𝑆𝑚3/𝑑

50 60,1 35,7 24,4 0,3 185 SS1

100 55,3 29,8 25,5 0,37 210 SS2

150 - - - - - Stable

200 - - - - - Stable

101

𝑄𝑜0 = 2000 𝑆𝑚3/𝑑

50 57,1 36,2 20,9 0,14 186 SS3

100 - - - - - Stable

150 - - - - - Stable

𝑄𝑜0 = 2500 𝑆𝑚3/𝑑

50 57,6 37,8 19,8 0,12 163 SS3

100 - - - - - Stable

150 - - - - - Stable

𝑄𝑜0 = 3000 𝑆𝑚3/𝑑

50 56,9 41,6 15,3 - 120 SS3

100 - - - - - Stable

150 - - - - - Stable *

* - flow exhibit irregular behavior; the value of period is either approximate or omitted.

102

Appendix E.

Table E.1. Riser base pressure readings with topside choking

Relative choke

opening,

[ - ]

Maximum observed

pressure,

[bara]

Minimum observed

pressure,

[bara]

Average

pressure,

[bara]

0,24 88,1 62,9 75,4

0,22 88,3 62,9 75,6

0,2 89,3 62,8 76,3

0,18 88,2 64,3 75,7

0,16 89,1 64,3 76,7

0,14 88,0 66,1 76,9

0,12 88,0 66,7 77,5

0,1 89,6 69,4 78,5

0,08 88,6 71,3 80,9

0,07 88,0 73,3 81,0

0,06 82,4 80,8 82,0

0,05 84,5 84,3 84,4

0,04 88,6 88,5 88,5

0,03 95,0 94,9 95,0